Challenge - can you understand this IAL text?

Qwynegold · Post by **Qwynegold** » Sat Jun 17, 2023 4:36 pm

I'll post a little statistics about the languages occurring in the texts. Here's about the first text:

More: show

Qwynegold · Post by **Qwynegold** » Sat Jun 17, 2023 4:39 pm

Hwhatting should theoretically be able to understand 93,2% of the first text.

Qwynegold · Post by **Qwynegold** » Sun Jun 18, 2023 4:59 am

And here comes the same statistics for the second text.

More: show

Qwynegold · Post by **Qwynegold** » Sun Jun 18, 2023 5:36 am

And for the third text:

More: show

Qwynegold · Post by **Qwynegold** » Sun Jun 18, 2023 10:32 am

I've made a graph to get a better overview of how many words each language has. The gray scale bars show how big a share each language had of the words in the three texts, with the black bar showing data for all three texts added together.

The magenta bars show how much of the entire dictionary (as it currently stands) each language makes up. If the gray scale bars go higher than the magenta bar, it means that that language is overrepresented in the texts.This will happen if the given language is represented among the most common words. If the gray scale bars are lower than the magenta bar, it means that the given language is underrepresented in the texts.

The yellow line marks the target for the languages that are written in lower case, ie. the bars (especially the magenta bars) should go above this line. (Classical Persian and Danish have no targets at all; they've just happened to get into the dictionary.) The green line marks the target for the languages that are written in all caps.

: Languages' share of vocabulary.png (119.6 KiB) Viewed 3876 times

So some languages don't hit the targets I've set up. I'm constantly trying to include these languages more when I create new words. And some languages shoot way past, see for example Italian. The problem with Italian is that if a word exists in Italian, at least one other Romance language will have an identical word. So whenever I borrow something from one Romance language, I often also get Italian as a bonus.

Qwynegold · Post by **Qwynegold** » Sun Jun 18, 2023 10:40 am

And here you can see which exact words come from which language.

Text 1

More: show

Text 2

More: show

Qwynegold · Post by **Qwynegold** » Sun Jun 18, 2023 10:42 am

Text 3

More: show

Ares Land · Post by **Ares Land** » Mon Jun 19, 2023 3:47 am

A lot of what I got from the second text is stuff people already guessed

More: show

Torco · Post by **Torco** » Mon Jun 19, 2023 3:32 pm

More: show

I think that the approach of having a set of languages one borrows from in order to make an IAL easy to understand prior to knowing it is probably only going to work if a) your set of natlangs is relatively small, and b) if they're related. it works for Interlingua (which I just unerstood effortlessly on first hearing, with only minor words giving me a "what?" moment), but it's not going to work for an IAL that borrows from classical persian, latin and oromo.

Of course, this begs the question just how close do the languages need to be for the results to be understandable a priori? at least we know that if you take the romance languages plus english it works, but i have a feeling it works too well... i.e. maybe you could do it with a broader set of languages while still being understandable. interlingua feels like it went the easy road lmao.

Qwynegold · Post by **Qwynegold** » Sat Jul 01, 2023 9:54 am

Ares Land wrote: ↑Mon Jun 19, 2023 3:47 am A lot of what I got from the second text is stuff people already guessed

More: show
Just figured out from the statistics cog means 'middle'
Tapa - hill
nama - people
omninama - everyone

mici and on are grammatical, though I'm not sure of the meaning.

These are all correct.

Qwynegold · Post by **Qwynegold** » Sat Jul 01, 2023 10:43 am

Torco wrote: ↑Mon Jun 19, 2023 3:32 pmthe padre nuestro thing is commonly used to test drive conlangs and, well, it was in verse.

Huh, I was unaware of this.

Torco wrote: ↑Mon Jun 19, 2023 3:32 pmI think that the approach of having a set of languages one borrows from in order to make an IAL easy to understand prior to knowing it is probably only going to work if a) your set of natlangs is relatively small, and b) if they're related. it works for Interlingua (which I just unerstood effortlessly on first hearing, with only minor words giving me a "what?" moment), but it's not going to work for an IAL that borrows from classical persian, latin and oromo.

Of course, this begs the question just how close do the languages need to be for the results to be understandable a priori? at least we know that if you take the romance languages plus english it works, but i have a feeling it works too well... i.e. maybe you could do it with a broader set of languages while still being understandable. interlingua feels like it went the easy road lmao.

Yeah, the difficulty here is the number of languages. I wanted to have representation from all over the world and not just European languages + Chinese. If all the languages here had an equal share, each language would only have 2.6% of the vocabulary. But in practice that isn't the case, because often times many languages have the same word. I just counted how many languages were the origin of each word in text 3, and got this result:

1 language: 140 - 59.8%
2 languages: 36 - 15.4%
3 languages: 23 - 9.8%
4 languages: 6 - 2.6%
5 languages: 19 - 8.1%
6 languages: 1 - 0.4%
8 languages: 1 - 0.4%
10 languages: 3 - 1.3%
16 languages: 2 - 0.9%
17 languages: 2 - 0.9%
19 languages: 1 - 0.4%

This means that the average word has 2.3 source languages. But there was still relatively little recognizability. I think part of the reason is the smallish phoneme inventory and restrictive phonotactics in Omni-kan. Very few of the English words got recognized in this test.

I wonder if the way I've created the vocabulary at least helps when learning glosses. But I don't have a way of testing that.

elgis · Post by **elgis** » Wed Jul 05, 2023 4:58 am

My attempt:

More: show

Some of the words look recognizable, but with so many source languages, I can never be sure about their meaning.

hwhatting · Post by **hwhatting** » Wed Jul 05, 2023 6:00 am

Qwynegold wrote: ↑Sat Jul 01, 2023 10:43 am Yeah, the difficulty here is the number of languages. I wanted to have representation from all over the world and not just European languages + Chinese. If all the languages here had an equal share, each language would only have 2.6% of the vocabulary. But in practice that isn't the case, because often times many languages have the same word. I just counted how many languages were the origin of each word in text 3, and got this result:
(snip)

This means that the average word has 2.3 source languages. But there was still relatively little recognizability. I think part of the reason is the smallish phoneme inventory and restrictive phonotactics in Omni-kan. Very few of the English words got recognized in this test.

I wonder if the way I've created the vocabulary at least helps when learning glosses. But I don't have a way of testing that.

I guess your approach has two problems: if each language contributes, say, about 10 words, and the languages are not similar to each other, you don't really help anybody much in learning your IAL. And if you simplify the phonology, you make it harder to recongnize words - it's a bit like the pük in Volapük is based on English "speak", but if you have to be told that in order to get it, that means it doesn't really help in learning the language.
OTOH, as the world won't use an IAL anyway, except if someone becomes world emperor and imposes it, you can treat it as an art project and don't need to concern yourself with learnability.

Moose-tache · Post by **Moose-tache** » Wed Jul 05, 2023 8:49 pm

As others have said, the texts were a lot more impenetrable than you seem to have expected, though context helped a lot.

I would argue that including words from Akan does not mean that Akan is "represented." This seems to miss the point of an IAL, which is to convince people to stop speaking Akan at meetings. English, too, of course. All of those languages exist for when the meeting is over and everyone goes to the bar. That's when Akan is actually represented, not in the IAL. If the purpose of the IAL is to be easy to understand, then it's not unfair or inappropriate to draw all of your vocabulary from a handful of mostly IE languages, if those languages are understood all over the world.

But that's all fine. My main criticism is, who the hell doesn't know who Jeffrey Henning is? If Richard Kenneway isn't your child's godfather, can you even call yourself a conlanger? Me and Hildegard Von Bingen used to do lines together.

Qwynegold · Post by **Qwynegold** » Sun Jul 23, 2023 12:42 pm

Finally I have some time to look at new posts carefully. Sorry for the long delay.

elgis wrote: ↑Wed Jul 05, 2023 4:58 am
More: show
Text 1:

- alam: known/knowledge ✗
- poste: after ✓
- mi: me/I ✓
- wakte: time ✓
- tapat: should/in front of/loyal/confide ✓ should
- wa: and ✓
- fahima: understand ✓
- omnis: all/every ✓
- iksakte: exactly ✓ exact

Text 2:

- piskos: fish ✗ the word for fish occurs somewhere, but it's not this word
- polus: sticks ✗
- kon: with ✓
- cigke: five/fifteen/fifty ✗
- mikros: small/tiny ✓
- postewakte: afterwards ✓

- wahit: one ✓
- wa: and ✓
- huwa: he ✗
- hiya: she ✗
- poste: after ✓
- koritore: corridor ✓
- tunel: tunnel ✓

Text 3:

- Nihogistan: Japan ✓
- Cogistan: China ✓
- tahun: years ✓
- tifisil: difficult ✓
- wino: wine ✓
- alkohol: alchohol ✓
- Sinto: Shinto ✓
- omniyalam: God ✗

Some of the words look recognizable, but with so many source languages, I can never be sure about their meaning.

You had good guesses. There were a few wrong guesses, but I think it might be a good idea to at least make a guess when you're not sure. Otherwise you won't get many correct answers either.

Qwynegold · Post by **Qwynegold** » Sun Jul 23, 2023 12:46 pm

hwhatting wrote: ↑Wed Jul 05, 2023 6:00 amit's a bit like the pük in Volapük is based on English "speak",

That's what it means?!!

hwhatting wrote: ↑Wed Jul 05, 2023 6:00 ambut if you have to be told that in order to get it, that means it doesn't really help in learning the language.

Ah, I see what you mean.

hwhatting wrote: ↑Wed Jul 05, 2023 6:00 amOTOH, as the world won't use an IAL anyway, except if someone becomes world emperor and imposes it, you can treat it as an art project and don't need to concern yourself with learnability.

Yes, that's basically what this is. I'm doing it for the fun and challenge of it. Though ideally it should be easily learnable as well.

Qwynegold · Post by **Qwynegold** » Sun Jul 23, 2023 12:56 pm

Moose-tache wrote: ↑Wed Jul 05, 2023 8:49 pmI would argue that including words from Akan does not mean that Akan is "represented." This seems to miss the point of an IAL, which is to convince people to stop speaking Akan at meetings. English, too, of course. All of those languages exist for when the meeting is over and everyone goes to the bar. That's when Akan is actually represented, not in the IAL.

Hmm, I'm not sure I can agree on that. Many IALs take their vocabulary from natlangs. This is not an engelang.

Moose-tache wrote: ↑Wed Jul 05, 2023 8:49 pmIf the purpose of the IAL is to be easy to understand, then it's not unfair or inappropriate to draw all of your vocabulary from a handful of mostly IE languages, if those languages are understood all over the world.

Ah, Esperanto has been criticized for being so Eurocentric, so that's why I wanted to try and represent more of the world. I have many design goals in this project, so maybe it's not completely like a prototypical IAL, but I think it's close enough to be called an IAL.

Moose-tache wrote: ↑Wed Jul 05, 2023 8:49 pmBut that's all fine. My main criticism is, who the hell doesn't know who Jeffrey Henning is? If Richard Kenneway isn't your child's godfather, can you even call yourself a conlanger? Me and Hildegard Von Bingen used to do lines together.

My contact with the rest of the conlanging community is like 95% through the ZBB, and I've never seen him mentioned here before.

keenir · Post by **keenir** » Mon Jul 24, 2023 2:20 am

Moose-tache wrote: ↑Wed Jul 05, 2023 8:49 pmBut that's all fine. My main criticism is, who the hell doesn't know who Jeffrey Henning is?

Even in the days of Langmaker, I thought Jeffrey Henning was a normal, average person - in the sense of making the occassional conlang and contributing its info to the website...like many other people back then. I didn't know it was his website until Zompist came out with the book Langmaker.

If Richard Kenneway isn't your child's godfather, can you even call yourself a conlanger?

Wait - who? I thought we were talking about Jeffrey Henning.

Me and Hildegard Von Bingen used to do lines together.

And when I hit my midlife crisis, I helped build the Hanging Gardens. Wait, are we not comparing ages?

well, thats what i used to tell people when they asked for my age.

xxx · Post by **xxx** » Mon Jul 24, 2023 3:46 am

Torco wrote: ↑Mon Jun 19, 2023 3:32 pm I think that the approach of having a set of languages one borrows from in order to make an IAL easy to understand prior to knowing it is probably only going to work if a) your set of natlangs is relatively small, and b) if they're related.

I totally agree, it's the only case where it works,
and in fact for it to work well you have to keep only one language,
that of the two speakers...

in our imaginary country,
where there are no financial or loyalty constraints,
there can be nothing at stake in legislating on the great figures of the national story...

keenir · Post by **keenir** » Mon Jul 24, 2023 4:43 am

xxx wrote: ↑Mon Jul 24, 2023 3:46 am
Torco wrote: ↑Mon Jun 19, 2023 3:32 pm I think that the approach of having a set of languages one borrows from in order to make an IAL easy to understand prior to knowing it is probably only going to work if a) your set of natlangs is relatively small, and b) if they're related.
I totally agree, it's the only case where it works,
and in fact for it to work well you have to keep only one language,
that of the two speakers...

in our imaginary country,
where there are no financial or loyalty constraints,

So...no debts and...does "no loyalty constraints" mean a person in that country can have as many loyalties to as many countries as they want - or they are only permitted no loyalties to anything but that country?

there can be nothing at stake in legislating on the great figures of the national story...

I don't understand: if there is nothing at stake, why make laws controlling the great figures? (laws like "no third terms for presidents anymore" or...?)

Zompist Bboard Again

Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?

Re: Challenge - can you understand this IAL text?