Vulgar: a language generator

A forum for all topics related to constructed languages
Linguistx
rupestrian
rupestrian
Posts: 9
Joined: 11 Apr 2017 11:45

Vulgar: a language generator

Post by Linguistx »

Hi. I've launched Vulgar. Vulgar auto-generates a usable conlang in the click on a button: a robust grammar and phonology outline, and a 2000 word vocabulary (with derivational words, homophones, etc.).

The goal was to build a tool that instantly creates a strong foundation for a conlang, while still leaving room to creatively flesh out the language.

I believe this this help people get over the hump of starting and abandoning projects because the beginning process is too time consuming.

The backend of the website is still very much under construction. There are many many more grammatical features I want to add, and probably a lot more on the vocabulary side.

I want your feedback and ideas for features!

If anyone is interested in purchasing the current premium version (gives you access to a 2000 word vocab and a custom orthography option) it's at a sale price of $19 via PayPal. Any purchase will give you access to all future updates via our email distribution list.

Features we're working on right now:

- Custom phonological rules
- Custom syllable structures
opipik
mayan
mayan
Posts: 1657
Joined: 12 Mar 2015 19:41

Re: Vulgar: a language generator

Post by opipik »

It's really good, but I think the default orthography is shit and too Englishy: /ʌ ɜ/ <u ur> and that's the only possibility, also horrible underdifferentiating of retroflex consonants and high vowels, /ø/ <u> which I don't know where did you get that
User avatar
Adarain
greek
greek
Posts: 511
Joined: 03 Jul 2015 15:36
Location: Switzerland, usually

Re: Vulgar: a language generator

Post by Adarain »

I've already given my criticisms over on reddit, so just so you don't have to hear the same thing again from others here, these are things I've already reported:
Spoiler:
* For some reason, this crashes my phone's browser after generating about 3-4 languages; a friend also told me that his Chrome crashed after a while.

Then about what it actually generates:

* Most phoneme inventories seem a bit odd. Most natlangs contain at least three plosives, but I don't think I've yet gotten a language with all of /p t k/.
* On the other hand, I also don't think I've yet seen a language with at least /ts/ or /tʂ/
* Generally speaking the consonant grid seems to be filled a bit randomly; it's also often quite empty. I've yet to see a language with anything that could be considered a large consonant inventory.
* Vowels seem fine.
* Every single language I've seen so far had:
-Definite and Indefinite articles
-Three-way tense distinction (past, present, future)
-One mood, one aspect
-A pronoun system that had exactly 3 persons and 2 numbers, sometimes with a distinction between masculine and feminine in the 3rd person singular (even if the nouns had no gender distinction; or more than 2 genders)
-If a language had noun classes, they were always labeled: Masculine and Feminine for the first two, Neuter for a third one, and then just enumeration (4th noun class).
All of these things could do with a lot more variation.
* I just generated a language which claims to have an "Ergative" and an "Absolutive". Yet, apart from the names given to the cases, there's nothing that tells me this is really the case: Word order is given with the terms "subject" and "object" (terms generally used for accusative languages) and the Absolutive was more marked: Image, which is _incredibly_ rare cross-linguistically. In the same language also, the [absolutive was used with prepositions see last column: Image, again this is unexpected. Basically I assume you just allowed "nominative" to sometimes be replaced by "ergative"; and then "accusative" by "absolutive". If you're just going to do such a substitution, please equate nominative and absolutive, they have much more in common.
* On formatting: I don't like the representation of the phoneme chart, personally. It's got too much blank space. Consider deleting empty rows and columns to clean the place up a bit. I'd look into how Glebshows its inventories, I find that much more pleasant to look at.
Last edited by Adarain on 12 Apr 2017 03:24, edited 1 time in total.
At kveldi skal dag lęyfa,
Konu es bręnnd es,
Mæki es ręyndr es,
Męy es gefin es,
Ís es yfir kømr,
Ǫl es drukkit es.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Vulgar: a language generator

Post by Creyeditor »

I really really like it, it's the best (only) thing like this that I know of. One thing that bothers me a lot is that most allophonic rules are so segment specific, i.e. they very rarely refer to natural classes of segments, both for the input and the context. It would be great to have more general rules, maybe something like Gleb has.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
mira
greek
greek
Posts: 759
Joined: 14 May 2016 11:59
Location: Switzerland
Contact:

Re: Vulgar: a language generator

Post by mira »

Adarain wrote:* For some reason, this crashes my phone's browser after generating about 3-4 languages; a friend also told me that his Chrome crashed after a while.
Sounds like a memory leak - hence why a computer with more RAM lasted longer than a puny phone. Remember to clear your lists and arrays Linguistix or people'll notice ;)
website | music | she/her | :gbr: native :deu: beginner
GrandPiano
mayan
mayan
Posts: 2080
Joined: 11 Jan 2015 23:22
Location: USA

Re: Vulgar: a language generator

Post by GrandPiano »

I agree with the other comments; I think this is really cool and I really like the idea, but at the same time, there's lots of room for improvement. In addition to what Adarain said, I think some of the allophony rules it creates are pretty strange; for example, one language I generated had "t → ŋ / a_".
User avatar
qwed117
mongolian
mongolian
Posts: 4094
Joined: 20 Nov 2014 02:27

Re: Vulgar: a language generator

Post by qwed117 »

GrandPiano wrote:I agree with the other comments; I think this is really cool and I really like the idea, but at the same time, there's lots of room for improvement. In addition to what Adarain said, I think some of the allophony rules it creates are pretty strange; for example, one language I generated had "t → ŋ / a_".
It isn't unexplainable; just weird. (And technically, isn't allophony)

You could understand it as deriving originally from a VtVn pattern that collapsed to atn where V1 = a (due to stress patterns), but collapsed to VtṼ> Vt, elsewhere. Still strange tho
Spoiler:
My minicity is [http://zyphrazia.myminicity.com/xml]Zyphrazia and [http://novland.myminicity.com/xml]Novland.

Minicity has fallen :(
The SqwedgePad
User avatar
All4Ɇn
mayan
mayan
Posts: 1765
Joined: 01 Mar 2014 07:19

Re: Vulgar: a language generator

Post by All4Ɇn »

Definitely agree with everyone else so far. Really cool but definitely could use some improvement. Like for starters the first language it autogenerated for me is "the language of Rápe" [xP]
GrandPiano
mayan
mayan
Posts: 2080
Joined: 11 Jan 2015 23:22
Location: USA

Re: Vulgar: a language generator

Post by GrandPiano »

qwed117 wrote:(And technically, isn't allophony)
How so? /t/ is allophonically realized as [ŋ] after /a/. The notation used is more often used for sound changes (and there's definitely some overlap between sound changes and allophony), but in context, I think it's pretty clear that it's meant to represent rules of allophony, unless I'm misunderstanding something.
User avatar
qwed117
mongolian
mongolian
Posts: 4094
Joined: 20 Nov 2014 02:27

Re: Vulgar: a language generator

Post by qwed117 »

GrandPiano wrote:
qwed117 wrote:(And technically, isn't allophony)
How so? /t/ is allophonically realized as [ŋ] after /a/. The notation used is more often used for sound changes (and there's definitely some overlap between sound changes and allophony), but in context, I think it's pretty clear that it's meant to represent rules of allophony, unless I'm misunderstanding something.
This is my understanding of allophony.
https://upload.wikimedia.org/wikipedia/ ... -chart.svg
Spoiler:
My minicity is [http://zyphrazia.myminicity.com/xml]Zyphrazia and [http://novland.myminicity.com/xml]Novland.

Minicity has fallen :(
The SqwedgePad
GrandPiano
mayan
mayan
Posts: 2080
Joined: 11 Jan 2015 23:22
Location: USA

Re: Vulgar: a language generator

Post by GrandPiano »

qwed117 wrote:
GrandPiano wrote:
qwed117 wrote:(And technically, isn't allophony)
How so? /t/ is allophonically realized as [ŋ] after /a/. The notation used is more often used for sound changes (and there's definitely some overlap between sound changes and allophony), but in context, I think it's pretty clear that it's meant to represent rules of allophony, unless I'm misunderstanding something.
This is my understanding of allophony.
https://upload.wikimedia.org/wikipedia/ ... -chart.svg
OK, that makes sense. I think it was still meant to be an allophone, even if it technically isn't, since otherwise it would have to be a diachronic sound change, which doesn't make sense in this context.

Since all of the specific commentary has been negative so far, one thing I like that I've noticed is that while the different forms of articles (this may apply to other aspects as well) seem to be randomly generated individually, there is still noticeable phonetic similarity between most of the forms, and usually some related forms are identical. This adds a little bit of extra realism and believability that wouldn't be there if the different forms had all been generated separately.
Linguistx
rupestrian
rupestrian
Posts: 9
Joined: 11 Apr 2017 11:45

Re: Vulgar: a language generator

Post by Linguistx »

opipik wrote:It's really good, but I think the default orthography is shit and too Englishy: /ʌ ɜ/ <u ur> and that's the only possibility, also horrible underdifferentiating of retroflex consonants and high vowels, /ø/ <u> which I don't know where did you get that
You can customise the orthography in the premium version (specifically because a non-English speaker requested it).

Also, while I agree that some of my default orthography choices are highly questionable (read: don't differentiate at all), I think it's fair to say that stranger things happens in natural languages. English is not un-notorious for this, and French uses the same letters for different vowels. Just two examples off the top of my head. Meanwhile, if I was going to map a unique letter to every IPA symbol, I'd just end up with the IPA, right?

Ok. So I could have made a function that looks at all the phonemes holistically and tries to do a nearest-to-possible job of differentiating all the sounds with a suite of diacritics. I considered it. But I also thought, why not throw in some things that just totally don't make sense? The aim of this tool is to make naturalistic languages, not IALs.

I would encourage you to check out the premium version if you really want to make it your own.
Adarain wrote:I've already given my criticisms over on reddit, so just so you don't have to hear the same thing again from others here, these are things I've already reported:
The phoneme inventories have been tweaked. I think they are quite a bit more naturalistic now, but tell me what you think. I'm not saying that something unnaturalistic won't happen, but also it's not the end of the world if it does.
The Ergative/absolutive thing has also been fixed. Thanks for pointing that out.

I'm going to crack into a suite of other grammatical ideas too.
Creyeditor wrote:I really really like it, it's the best (only) thing like this that I know of. One thing that bothers me a lot is that most allophonic rules are so segment specific, i.e. they very rarely refer to natural classes of segments, both for the input and the context. It would be great to have more general rules, maybe something like Gleb has.
Yeah that's not a bad idea. I'm taking all these suggestions down. Just a matter of when I tackle them.
OTʜᴇB wrote:Sounds like a memory leak - hence why a computer with more RAM lasted longer than a puny phone. Remember to clear your lists and arrays Linguistix or people'll notice ;)
Possibly, I'll definitely look into it. I have a suspicion that certain seed numbers (seed number that are randomly generated when the user leaves the seed field blank) crash the program too. I at least know that 0 and 0.000000000000003 crash the script (don't try it).
GrandPiano wrote:I agree with the other comments; I think this is really cool and I really like the idea, but at the same time, there's lots of room for improvement. In addition to what Adarain said, I think some of the allophony rules it creates are pretty strange; for example, one language I generated had "t → ŋ / a_".
I will just quickly say that all the phonological rules were taken from Index Diachronica.
qwed117 wrote:
GrandPiano wrote:
qwed117 wrote:(And technically, isn't allophony)
How so? /t/ is allophonically realized as [ŋ] after /a/. The notation used is more often used for sound changes (and there's definitely some overlap between sound changes and allophony), but in context, I think it's pretty clear that it's meant to represent rules of allophony, unless I'm misunderstanding something.
This is my understanding of allophony.
https://upload.wikimedia.org/wikipedia/ ... -chart.svg
That's interesting. That was actually a knowledge gap for me.
Linguistx
rupestrian
rupestrian
Posts: 9
Joined: 11 Apr 2017 11:45

Re: Vulgar: a language generator

Post by Linguistx »

GrandPiano wrote:Since all of the specific commentary has been negative so far, one thing I like that I've noticed is that while the different forms of articles (this may apply to other aspects as well) seem to be randomly generated individually, there is still noticeable phonetic similarity between most of the forms, and usually some related forms are identical. This adds a little bit of extra realism and believability that wouldn't be there if the different forms had all been generated separately.
Thanks! Yes, this was obviously quite deliberate. Negative feedback is fine haha. It shows me what people care about.
User avatar
DesEsseintes
mongolian
mongolian
Posts: 4331
Joined: 31 Mar 2013 13:16

Re: Vulgar: a language generator

Post by DesEsseintes »

qwed117 wrote:
GrandPiano wrote:I agree with the other comments; I think this is really cool and I really like the idea, but at the same time, there's lots of room for improvement. In addition to what Adarain said, I think some of the allophony rules it creates are pretty strange; for example, one language I generated had "t → ŋ / a_".
It isn't unexplainable; just weird. (And technically, isn't allophony)
You could understand it as deriving originally from a VtVn pattern that collapsed to atn where V1 = a (due to stress patterns), but collapsed to VtṼ> Vt, elsewhere. Still strange tho
Here's another possible(?) explanation:

/t/ debuccalises to [ʔ] before /a/ (a very likely environment as /a/ is the lowest/openest vowel), and then rhinoglottophilia kicks in to turn [ʔ] into [ŋ]. Voilà, two simple and widely attested processes to explain this rather than a complex reconstruction. [B)]
User avatar
Ahzoh
mongolian
mongolian
Posts: 4191
Joined: 20 Oct 2013 02:57
Location: Canada

Re: Vulgar: a language generator

Post by Ahzoh »

Linguistx wrote:The aim of this tool is to make naturalistic languages, not IALs.
Shallow orthography (diacritic'd letters; c.f. a deep orthography like English) does not an IAL make. And naturalistic is not every seed written with English-style orthography.
Image Śād Warḫallun (Vrkhazhian) [ WIKI | CWS ]
User avatar
Adarain
greek
greek
Posts: 511
Joined: 03 Jul 2015 15:36
Location: Switzerland, usually

Re: Vulgar: a language generator

Post by Adarain »

Linguistx wrote:
Adarain wrote:I've already given my criticisms over on reddit, so just so you don't have to hear the same thing again from others here, these are things I've already reported:
The phoneme inventories have been tweaked. I think they are quite a bit more naturalistic now, but tell me what you think. I'm not saying that something unnaturalistic won't happen, but also it's not the end of the world if it does.
The Ergative/absolutive thing has also been fixed. Thanks for pointing that out.

I'm going to crack into a suite of other grammatical ideas too.
I will say the phoneme inventories look much better now. However, they're still boring. There's nearly no variation in size, no crazy 80+ phoneme inventories (nor even much more sensible stuff like say Irish or Russian, which take an average sized consonant inventory but double its size because they add palatalization). The only contrast I ever see is one of voicing. I kinda assume this has to do with your bug concerning diacritics though.
At kveldi skal dag lęyfa,
Konu es bręnnd es,
Mæki es ręyndr es,
Męy es gefin es,
Ís es yfir kømr,
Ǫl es drukkit es.
Linguistx
rupestrian
rupestrian
Posts: 9
Joined: 11 Apr 2017 11:45

Re: Vulgar: a language generator

Post by Linguistx »

Ahzoh wrote:
Linguistx wrote:The aim of this tool is to make naturalistic languages, not IALs.
Shallow orthography (diacritic'd letters; c.f. a deep orthography like English) does not an IAL make. And naturalistic is not every seed written with English-style orthography.
Sure, you're right. I mean I could encourage you to buy the full version, however the custom orthography option is basically just a find-replace tool in the order you write the rules. But you can do things like:

q > k
ka > qa

Making every 'q' a 'k' except before an 'a'. But that's not "deep orthography", it's just applying rules in order of application. Alternatively there's also an option the fuck the orthography off completely, leaving only IPA and definition.

I do like the idea of deep orthography now that you mention it, but I would probably make it a premium feature if and when I got around to it.
Linguistx
rupestrian
rupestrian
Posts: 9
Joined: 11 Apr 2017 11:45

Re: Vulgar: a language generator

Post by Linguistx »

Adarain wrote:
Linguistx wrote:
Adarain wrote:I've already given my criticisms over on reddit, so just so you don't have to hear the same thing again from others here, these are things I've already reported:
The phoneme inventories have been tweaked. I think they are quite a bit more naturalistic now, but tell me what you think. I'm not saying that something unnaturalistic won't happen, but also it's not the end of the world if it does.
The Ergative/absolutive thing has also been fixed. Thanks for pointing that out.

I'm going to crack into a suite of other grammatical ideas too.
I will say the phoneme inventories look much better now. However, they're still boring. There's nearly no variation in size, no crazy 80+ phoneme inventories (nor even much more sensible stuff like say Irish or Russian, which take an average sized consonant inventory but double its size because they add palatalization). The only contrast I ever see is one of voicing. I kinda assume this has to do with your bug concerning diacritics though.
Yeah I think there's still room for improvement. I've got to stop replying to comments and actually work.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Vulgar: a language generator

Post by Creyeditor »

Wow, I really like the latest version of this [:)]
It's getting better and better,
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
GamerGeek
sinic
sinic
Posts: 345
Joined: 17 May 2017 18:10
Location: The Universe
Contact:

Re: Vulgar: a language generator

Post by GamerGeek »

I don't like the fact that it has a premium version, but it's still really cool.
Post Reply