Omni-kan syntax

[Rounin Ryuuji]

All this said, this being your language, you ought to do with it what you like, but I think w is, unfortunately, the most logical choice. I rather like it as a letter, however. It is not perhaps as pretty as romance v, but it does have a pleasant, Germanic-Japonic homeliness to it.

[Qwynegold]

Okay, I'll keep the W.

Is all of this adjusted for number of speakers?

Good question. No, that's the number of individual languages. It would be impossible to calculate the number of speakers for any feature. But I'll have a look and make a guesstimation based on the big languages...

[Qwynegold]

All this said, this being your language, you ought to do with it what you like, but I think w is, unfortunately, the most logical choice.

Yeah, I know. It's hard not to treat it like an artlang.

[Rounin Ryuuji]

Is all of this adjusted for number of speakers?

Good question. No, that's the number of individual languages. It would be impossible to calculate the number of speakers for any feature. But I'll have a look and make a guesstimation based on the big languages...

The most common Indo-European languages seem to be Spanish, English, and Hindustani (referencing Greek, Latin, and Sanskrit, and in some varieties Persian or Arabic, for classical vocabularies); if it weren't for Mandarin being such a very common language, too, I would think you might have a structural beginning with that Indo-European bunch.

Yeah, I know. It's hard not to treat it like an artlang.

If we're honest with ourselves, auxiliary languages probably aren't going to catch on for a good, long while, so it's probably fine to be at least a little artistic somewhere. Bearing in mind I'm a great fan of deep, historically-conditioned orthographies.

[Qwynegold]

I actually did the numbers.

Code: Select all

      L1 speakers   L2 speakers
Gen-N 1,954,342,550 2,403,495,970
N-Gen 1,425,775,327 2,577,478,800

This includes all my focus languages, except for some that lacked data points in WALS. English is also excluded because English has both orders. One has to take the figures for number of speakers with a grain of salt though. Which order wins depends on whether you count L2 speakers or not. And the numbers for L2 are probably even less reliable than for L1... Also, I've discovered another relevant issue, see below.

I was trying to come up with what to do about relative clauses, and was thinking of some sentences. Then I came upon this sentence: Everything that we said, are just words of yesterday. One attempt at Omni-kanizing it is things all be words GEN yesterday only, we say they.

A question that arises is: how should an adjective (only in the example) be placed when there's a genitive also involved? If I place it like above, then it seems like only is modifying yesterday instead of words. I guess I could do "words only GEN yesterday", though wouldn't it be more elegant if it was instead "yesterday GEN words only"? Because as it is now, everything is getting packed to the right of the noun, so it would be a little less messy if one thing went to the left.

Anyhow, let's ignore the word "only" for now, and look at the order of the two clauses. When it's done like two separate clauses like this, instead of a relative clause nested inside the matrix clause, I would for some reason like to put the "less important" information first: we say things, that all be words GEN yesterday. Maybe you could choose whether you want to put the relative clause first or last. Well, if you can even say that there is a relative clause. It's not modifying anything, right?

So then you use an anaphor in the latter clause, whichever that is. Maybe this could work: If it refers to the subject of the previous clause, use a pronoun. If it refers to anything else, use that, or a repetition of that noun, or a repetition of that noun with that. I'll have to try some sentences to see if confusion would arise about what the antecedent is when it's not the subject...

[Travis B.]

A question that arises is: how should an adjective (only in the example) be placed when there's a genitive also involved? If I place it like above, then it seems like only is modifying yesterday instead of words. I guess I could do "words only GEN yesterday", though wouldn't it be more elegant if it was instead "yesterday GEN words only"? Because as it is now, everything is getting packed to the right of the noun, so it would be a little less messy if one thing went to the left.

That is a major question in my IAL, because in it there are no adjectives per se, just stative verbs, so the equivalent of attributive usage of adjectives are relative clauses containing stative verbs. Likewise, there are no adpositions, but so instead of attributive adpositions there are rather (often stative) transitive verbs in relative clauses. Hence it is very easy to stack up multiple relative clauses qualifying the same noun. My solution is to put the intransitive relative clauses before the transitive ones, to reduce the chance of a relative clause being misunderstood as qualifying a noun in another relative clause rather than in the original noun being qualified.

[Qwynegold]

That's the one big difference between our IALs. Yours uses relative clauses for everything, and everything is a verb. ^_^;

Maybe I should translate some things to see how it feels to have which order of genitive and noun.

[Qwynegold]

So then you use an anaphor in the latter clause, whichever that is. Maybe this could work: If it refers to the subject of the previous clause, use a pronoun. If it refers to anything else, use that, or a repetition of that noun, or a repetition of that noun with that. I'll have to try some sentences to see if confusion would arise about what the antecedent is when it's not the subject...

I just tried a few sentences real quick. I will maybe want to switch around the anaphoric pronoun and "that". And I will probably allow the two clauses to be placed in either order. I should try translating an entire page of a book...

[Qwynegold]

Some notes on verbal stuff. These might change a bit.

Tense
No tense marking is used. Tense can be implied through time references, e.g. “yesterday”, “next month”, “in 1990”, “when I start school”, “every evening”, or “that time”, which refers to some point in time already mentioned earlier. “Now” can be used to specify present tense.

Perfect aspect
Expressed with “already”. If the sentence has a time reference, it's either past perfect or future perfect depending on the time reference. Without any time reference, it's present perfect.

Perfective aspect
Expressed with “complete”. Used about actions that have been finished and completed. Can't be used for actions that have no natural completion point.

Continuous or iterative aspect
Expressed through reduplication. When used on compound verbs, each word is repeated separately, like word1 verb2 → word1 word1 verb2 verb2. (Auxiliaries are not repeated.) Used for actions that are being repeated continuously, or which are ongoing. If there is a time reference, it's either past or future depending on the reference. If there is no time reference, then it's present.

Other auxiliaries
“Quit” used when one stops an action without completing it and without intending to resume it. “Stop” is used when one stopped an action but resumed it later, or intends to resume it. “Should”, “must”, “can”...

I'm not sure if I should have so many verbal distinctions, especially the quit/stop distinction. People would likely mix these up. On the other hand, most of this stuff is expressed through verbs with transparent meanings, so it shouldn't be that hard to understand how they're used.

[Qwynegold]

Just a quick little thing I've been thinking about... Like in "I like it" and please in "it pleases me" mean the same thing, they just have different syntax. Which of these two verbs do you think is more common world wide, if you understand what I mean?

[Rounin Ryuuji]

Both English and French use both with different words, cf. the example I like it v. It pleases me, and also Celà me plaît v. Je l'aime, though the former seems to be more common; Spanish also does me gusta; Japanese does a weird Japanese thing with an adjective suki(na), which, if I've understood right, is a contraction of an earlier pharse suki naru, a contraction of earlier suki ni aru, which is roughly (whatever) is fine/good by me or (whatever) is nice/good. Maybe a copular phrase meaning "(whatever) is good" would prove fairly universally comprehensible (though this could also, I admit, be interpreted to mean "useful"); a phrase "(whatever) + (copula) + nice/enjoyable" might also work?

[bradrn]

Just a quick little thing I've been thinking about... Like in "I like it" and please in "it pleases me" mean the same thing, they just have different syntax. Which of these two verbs do you think is more common world wide, if you understand what I mean?

From Dixon’s Basic Linguistic Theory:

Verbs describing mental feeling are of particular interest … either Experiencer or Stimulus could hold major responsibility for the
state of mind. Indeed, English has two semantic types involving these roles. Verbs of the LIKING type focus on the Experiencer as relating to the success of this mental state. … In contrast, verbs of the ANNOYING type (including offend, anger, please, impress, and entertain) have Stimulus in A role and Experiencer as O. Other LIKING/ANNOYING verb pairs—in which associations between Experiencer and Stimulus roles are interchanged between members of the pair—include like/please, admire/impress, and fear/terrify.

Many languages have some LIKING verbs which are transitive, with the Experiencer role as A and Stimulus as O. Many fewer have verbs like those of
the ANNOYING type in English, with role/function linkings reversed.

That is, verbs like ‘like’ are far more common than verbs like ‘please’.

[Qwynegold]

Ryuuji: Thanks for the suggestion. I'll have to think about if I want to use a construction of that sort instead of a simple verb.

That is, verbs like ‘like’ are far more common than verbs like ‘please’.

OMG I didn't think there would actually be an answer that clear. This really makes me want to go the easy route of like. But I guess I have to consider other feeling verbs as well, and make everything work in the same way.

[bradrn]

That is, verbs like ‘like’ are far more common than verbs like ‘please’.

OMG I didn't think there would actually be an answer that clear. This really makes me want to go the easy route of like. But I guess I have to consider other feeling verbs as well, and make everything work in the same way.

The general rule is that the most animate participant gets assigned to A. If you’re interested in the details, I highly recommend seeing if you can obtain a copy of Basic Linguistic Theory vol. 2: it has an excellent chapter about transitivity which covers this stuff in great detail.

Another approach, possibly more suitable for an IAL, is to get rid of transitivity altogether and simply case-mark each NP based on its semantic role — ‘agent’, ‘undergoer’, ‘experiencer’, ‘recipient’ etc. This means that you don’t have quite so many problems with assigning participants to A and O, though on the other hand this approach is rarer amongst natlangs.

[Qwynegold]

The general rule is that the most animate participant gets assigned to A.

Ah, that makes sense.

Another approach, possibly more suitable for an IAL, is to get rid of transitivity altogether and simply case-mark each NP based on its semantic role — ‘agent’, ‘undergoer’, ‘experiencer’, ‘recipient’ etc. This means that you don’t have quite so many problems with assigning participants to A and O, though on the other hand this approach is rarer amongst natlangs.

Mm, nah. I'm not marking anything in this language, because it seems to be unintuitive to people. Cf. pidgins and very non-fluent speech of any synthetic language. Transitivity is a sticky issue though; people may have different ideas about which verbs are transitive and which are intransitive.

[Qwynegold]

I was "translating" some text yesterday, and realized that my current ideas about word order are not working. The problem was that I had head final order in compounds and head first order in mostly everything else. It turns out there isn't actually much difference between compounds and words being modified by separate words. So oftentimes it was quite ambiguous as to which order I should use for a specific group of words. I'll have to do some more "translation" and try different things...

[Qwynegold]

Earlier I compiled data to find out which order of genitive and noun was most common, based on number of speakers. This included all languages with over 50 million speakers, plus a few others with over 10 million speakers. The results of that weren't very conclusive. But now I did the same with adjective and noun, demonstrative and noun, and numeral and noun. The results of this were that final headedness was clearly more common, regardless of whether you looked at all languages or just SVO-langs, and whether you included L2 speakers or not. So I guess I'm going with Adj-N, Dem-N and Num-N. But I'm probably going with the "of" type of genitive. I tried mock-translating some text, and it came out very similar to English.

I finally posted that survey, and I'm pleased to see that I've already gotten a few answers.

[Qwynegold]

Does it make sense to distinguish between "that" and "it"? Most languages don't seem to make the distinction. But I'm worrying about the massive use of "that" that this conlang will have.

[bradrn]

Does it make sense to distinguish between "that" and "it"? Most languages don't seem to make the distinction. But I'm worrying about the massive use of "that" that this conlang will have.

Where did you see this claim? Bhat (2004) confirms it, but it’s hardly ‘most languages’: ‘That study [sampling 255 languages from WALS] indicated that two-person languages [i.e. ‘that’=‘it’] slightly outnumber three-person languages [i.e. ‘that’≠‘it’]: It appears to be possible to regard 126 of these languages as two-person languages, and only 99 as three-person languages.’

But in the same book, he mentions that many three-person languages still distinguish between pronouns and demonstratives. One option you may be interested in is for them to be related derivationally, for instance as in Asheninca:

pronouns-demonstratives-asheninca.png (9.89 KiB) Viewed 8342 times

(Sorry about the image, but I didn’t feel like figuring out how to merge cells in a BBCode table.)

Khasi has a slightly different system: the demonstratives are primary, and you get pronouns by adding kaṛ ‘person’ to them.

[Vardelm]

(Sorry about the image, but I didn’t feel like figuring out how to merge cells in a BBCode table.)

I don't believe you can, so you are excused.