(Conlangs) Q&A Thread - Quick questions go here

A forum for all topics related to constructed languages
User avatar
Vlürch
greek
greek
Posts: 452
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Vlürch »

Solarius wrote: 06 Jul 2022 17:32 Debating the romanization for a the palatal series of a language I'm working on. The current romanization is /t͡ɕ ɲ ʎ j/<tj nj lj y> because I liked the odd and clunky look of the Cj digraphs, but that clunky look makes a lot of words look sort of strange, especially since <i l> are pretty common too, giving it a weird homogeneous appearance. I also don't like the inconsistency of using <Cj> when there's no <j> in the language.

I'm considering either:
/t͡ɕ ɲ ʎ j/<j ñ ll y>
or
/t͡ɕ ɲ ʎ j/<c ñ y j>

Digraphs are totally fine (the syllable structure prevents basically all consonant clusters), so it's really just a measure of phonaesthetics. What do y'all think?
Well, it depends on a number of factors, doesn't it?

1) If set in the "real world" or an althistory, where is the language spoken? What languages could its orthography have been based on/influenced by?
/t͡ɕ ɲ ʎ j/<j ñ ll y> has a pretty Spanish vibe, possibly hinting at historical contact with Spanish, while /t͡ɕ ɲ ʎ j/<c ñ y j> is pretty unique AFAIK and would make me think of the language's spelling having been reformed/standardised recently. The former would make sense if the language had (historical) contact with Spanish, if whoever/whichever committee conically devised the orthography perceived Spanish as prestigious and worth emulating, etc. On the other hand, the latter could be indicative of the conical orthographists having decided "we'll do something practical" without consulting anyone else, which could be divisive among linguists, language nerds and/or the language's speakers themselves (see eg. the ongoing Kazakh orthographic reform), and competing orthographies might also exist; AFAIK this kind of situation is fairly common with American indigenous languages.

2) Is the orthography supposed to be "naturalistic" (whatever that means for an orthography; obviously ANADEW always applies) or not?
As in, did the orthography arise gradually through historical spelling reforms or was it recently revamped and modelled after a different language's orthography, or was it pulled out of thin air by a field linguist as an ad hoc system that then spread among the speakers? Things like this can make a huge difference for what an orthography "should" be like. Obviously you can handwave anything easily, but a language whose orthography gradually developed to its current form through various spelling reforms is more likely to be more "reserved", since they probably wouldn't just randomly for no reason decide that /t͡ɕ ɲ ʎ j/, written as <j ñ ll y> for centuries, would be written as <tsch nghy lghy ı> from now on. I mean, it can happen, but... on the other hand, if the orthography wasn't standardised until recently, especially if the language was historically mostly unwritten, it could be whatever. They'd probably model it after some other orthography to some degree, likely a language perceived as prestigious, but not necessarily; if /t͡ɕ ɲ ʎ j/ had been written in a dozen different and inconsistent ways, there's less incentive to "maintain a tradition" because a tradition wouldn't even exist.

3) What kind of orthography do you think is most fitting, purely personally?
I mean, we all have our own phonaesthetic associations stemming from both purely personal and cultural factors. Do you want to evoke a certain abstract vibe through the orthography, and if so, what kind of vibe is that? Do you want to intentionally mismatch the vibe you get from the phonology with an entirely different vibe from the orthography? If for example <j ñ ll y> make you think "elegant" and <c ñ y j> make you think "cute", what does the phonology altogether make you think? What adjectives would you use to describe the language's sound? Is it "elegant", "cute", or the dreaded "guttural", or...?

4) What is the rest of the orthography like?
If you have /t͡ɕ ɲ ʎ j/ contrasting with /t͡s n l ɹ/ for example, it's worth considering if the former would be represented as "paired" with or "variants" of the latter. You know, how a bunch of languages have /t͡ʃ/ as <č> and /t͡s/ as <c> and whatnot. With /j/ and /ɹ/ it might be less likely to follow that kind of pattern since they're likely perceived as a semivowel and rhotic (but not necessarily), but with /t͡ɕ ɲ ʎ/ and /t͡s n l/ most languages would make some kind of orthographic link between them. So, if you have /t͡s n l/ <ts n l>, then something like <tš ň ľ> or <tś ń ĺ> or <tsy ny ly> or whatever for /t͡ɕ ɲ ʎ/ would make sense. If they don't contrast with alveolar (or velar, or...) consonants that they could be perceived as "paired" with or "variants" of, then they'd be more likely represented by "simpler" letters. But again, this kind of things are pretty arbitrary and not necessarily entirely "consistent" and those kinds of associations may well arise later from the spelling rather than the other way, so you can handwave pretty much anything.

5) What is the rest of the phonology like?
You said it has no consonant clusters, but what about other considerations? If you do have /t͡s n l ɹ/ or something as well, are they and /t͡ɕ ɲ ʎ j/ differentiated in all contexts or do they (or some of them) merge in some environments, eg. before /i/? If they do merge, is the orthrography more phonemic or phonetic? As in, if it has eg. /t͡s n l/ <z n l> and /t͡ɕ ɲ ʎ/ <ch ñ ll> modelled after Spanish and they merge before /i/, would [t͡ɕi ɲi ʎi] be written <zi ni li> or <chi ñi lli>? If you use consistent digraphs, having eg. /t͡s n l/ <c n l> /t͡ɕ ɲ ʎ/ <cy ny ly>, and they merge before /i/, then writing [t͡ɕi ɲi ʎi] as just <ci ni li> would make sense; if there are no diphthongs or consecutive vowels, you could well also have /t͡ɕ ɲ ʎ/ written <ci ni li> to begin with. Or would it be etymological, as in they didn't merge until after the orthrography was established, and would still be differentiated in writing? And if they don't merge, then would eg. /t͡ɕ ɲ ʎ/ <ci ni li> be unaesthetic because you'd end up with [t͡ɕi ɲi ʎi] <cii nii lii>? And so on.

Of course I didn't say anything deeply insightful or anything, but the point is that orthographies are best considered as a whole and together with the phonology.
User avatar
aliensdrinktea
hieroglyphic
hieroglyphic
Posts: 69
Joined: 11 Oct 2020 17:28

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by aliensdrinktea »

Is secondary stress necessarily weight-sensitive if primary stress is? The system I have in mind is that primary stress falls on the first syllable, unless that syllable is light and the following one is heavy, in which case the second syllable is stressed. Secondary stress falls on every alternate syllable after that regardless of weight (though final syllables can only be stressed if heavy). Is this naturalistic?
User avatar
Omzinesý
mongolian
mongolian
Posts: 4081
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Omzinesý »

aliensdrinktea wrote: 26 Jul 2022 01:25 Is secondary stress necessarily weight-sensitive if primary stress is? The system I have in mind is that primary stress falls on the first syllable, unless that syllable is light and the following one is heavy, in which case the second syllable is stressed. Secondary stress falls on every alternate syllable after that regardless of weight (though final syllables can only be stressed if heavy). Is this naturalistic?
I'm quite sure there is no problem with your system.
But somebody with more knowledge can give a more sophisticated answer.
My meta-thread: viewtopic.php?f=6&t=5760
User avatar
sangi39
moderator
moderator
Posts: 3024
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by sangi39 »

aliensdrinktea wrote: 26 Jul 2022 01:25 Is secondary stress necessarily weight-sensitive if primary stress is? The system I have in mind is that primary stress falls on the first syllable, unless that syllable is light and the following one is heavy, in which case the second syllable is stressed. Secondary stress falls on every alternate syllable after that regardless of weight (though final syllables can only be stressed if heavy). Is this naturalistic?
I use a similar system in one of my conlangs, and I swear I've seen it appear in natlangs, so you're probably fine if you're going for naturalism [:)]
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.
User avatar
aliensdrinktea
hieroglyphic
hieroglyphic
Posts: 69
Joined: 11 Oct 2020 17:28

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by aliensdrinktea »

Awesome, thanks!
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by eldin raigmore »

aliensdrinktea wrote: 26 Jul 2022 01:25 Is secondary stress necessarily weight-sensitive if primary stress is? The system I have in mind is that primary stress falls on the first syllable, unless that syllable is light and the following one is heavy, in which case the second syllable is stressed. Secondary stress falls on every alternate syllable after that regardless of weight (though final syllables can only be stressed if heavy). Is this naturalistic?
According to various stress-type databases available online,
Either one, or both, or neither, of primary stress and secondary stress, can be weight-sensitive, whether or not the other is.
Also: if both are weight-sensitive, they may have different ways (weighs?) of determining syllable-weight. (Or they may use the same weights, of course.)

…..

The system you describe sounds naturalistic to me. If a natlang has that system you may be able to look it up on one of those databases.
User avatar
aliensdrinktea
hieroglyphic
hieroglyphic
Posts: 69
Joined: 11 Oct 2020 17:28

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by aliensdrinktea »

eldin raigmore wrote: 26 Jul 2022 18:04 According to various stress-type databases available online,
Either one, or both, or neither, of primary stress and secondary stress, can be weight-sensitive, whether or not the other is.
Also: if both are weight-sensitive, they may have different ways (weighs?) of determining syllable-weight. (Or they may use the same weights, of course.)

…..

The system you describe sounds naturalistic to me. If a natlang has that system you may be able to look it up on one of those databases.
That’s really interesting! Could you give me a link to one of these databases?
Salmoneus
MVP
MVP
Posts: 3030
Joined: 19 Sep 2011 19:37

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Salmoneus »

In WALS' sample, 50% of languages with weight-sensitive stress for either primary or secondary stress have insensitive stress for the other (33% weighted primary unweighted secondary, 17% vice versa).

To be fair, I expect that a large percentage of those cases will be where one type of stress just has an extremely simple rule where secondary stress is just unrelated to primary stress (like "secondary stress is always on the initial syllable" or "all long vowels take secondary stress"), but I would expect at least some would have more complicated rules like yours.


One concern may be that because your primary stress may be on either the first or second syllable, and your secondary stress iterates on every second syllable after that, your language will have both trochaic and iambic rhythm. WALS claims that this is extremely rare (only 4 examples). However, their approach to this appears to be extremely theory-heavy (they mention that some people outright refuse to believe that iambs exist, which is... a theory i suppose), and I don't have any idea what they really mean (eg they characterise English as unproblematically solely trochaic, which is is... another theory, I guess....). So you might want to look into this but I wouldn't automatically panic because of it.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Creyeditor »

aliensdrinktea wrote: 26 Jul 2022 01:25 Is secondary stress necessarily weight-sensitive if primary stress is? The system I have in mind is that primary stress falls on the first syllable, unless that syllable is light and the following one is heavy, in which case the second syllable is stressed. Secondary stress falls on every alternate syllable after that regardless of weight (though final syllables can only be stressed if heavy). Is this naturalistic?
I think it might make sense to schematically write down the stress patterns you predict for words of some length. I use O for any syllable, H for heavy ones and L for light ones. ' marks primary stress, . is a syllable boundary and , is secondary stress. This might make it easier to compare to other languages.

2-syllables
'L.L
L.'H
'H.L
'H.H

5-syllables
'L.L.,O.O.L
L.'H.O.,O.L
'H.L.,O.O.L
'H.H.,O.O.L
'L.L.,O.O.,H
L.'H.O.,O.H
'H.L.,O.O.,H
'H.H.,O.O.,H


If this is correct, one analytical tool that you might want to look at is extrametricality. Some syllables are ignored for the sake of stress. For your system it looks like a final light syllable is ignored, as well as an initial light syllable followed directly by a heavy syllable. These could be considered (preferably) extrametrical. This might solve the foot problem Sal mentioned, as this tool was also used by the WALS people, IIRC.
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
aliensdrinktea
hieroglyphic
hieroglyphic
Posts: 69
Joined: 11 Oct 2020 17:28

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by aliensdrinktea »

Creyeditor wrote: 26 Jul 2022 21:31 If this is correct, one analytical tool that you might want to look at is extrametricality. Some syllables are ignored for the sake of stress. For your system it looks like a final light syllable is ignored, as well as an initial light syllable followed directly by a heavy syllable. These could be considered (preferably) extrametrical. This might solve the foot problem Sal mentioned, as this tool was also used by the WALS people, IIRC.
So I did some research, and it turns out that extrametrical moras are a thing, and that a final mora being treated as extrametrical produces alternation between stressed heavy/unstressed light final syllables. So that's promising.

On the other hand, initial extrametricality (be it initial moras, syllables, or feet) seems to be controversial and rarely if ever attested.

BUT, I've been messing around with different secondary stress patterns on a whiteboard and came up with a system that I like better aesthetically—plus it comes without the theoretical problems!

1. Primary stress falls on the first syllable, unless that syllable is light and the second is heavy, in which case the second syllable receives primary stress. (Same rule as before, stated by WALS to be the most common in languages with left-edge bounded stress.)
2. Trochaic feet are constructed from right to left, ignoring syllable weight.
3. Secondary stress is assigned to the head of each foot, so long as it does not clash with the primary stress.
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by eldin raigmore »

aliensdrinktea wrote: 26 Jul 2022 18:26 That’s really interesting! Could you give me a link to one of these databases?
This is one:
http://st2.ullet.net/?content=search_pattern&clear=1
Guidance: http://st2.ullet.net/?content=page&id=17
If memory serves, you can also search for secondary-stress patterns, but I don’t know where it tells you how to do that.


You should probably try at least two databases, just in case.
I used to look at https://udel.edu/~amandapa/stress.pdf. But that might be a dead link now.
User avatar
aliensdrinktea
hieroglyphic
hieroglyphic
Posts: 69
Joined: 11 Oct 2020 17:28

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by aliensdrinktea »

eldin raigmore wrote: 27 Jul 2022 06:23
aliensdrinktea wrote: 26 Jul 2022 18:26 That’s really interesting! Could you give me a link to one of these databases?
This is one:
http://st2.ullet.net/?content=search_pattern&clear=1
Guidance: http://st2.ullet.net/?content=page&id=17
If memory serves, you can also search for secondary-stress patterns, but I don’t know where it tells you how to do that.


You should probably try at least two databases, just in case.
I used to look at https://udel.edu/~amandapa/stress.pdf. But that might be a dead link now.
The second link is dead, but thank you! I'll be saving the first!
Salmoneus
MVP
MVP
Posts: 3030
Joined: 19 Sep 2011 19:37

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Salmoneus »

The problem with extrametricality is precisely that it allows any actual linguistic data to "confirm" literally any proposed "rule" - just edit out all the bits that don't fit your rule.

For instance, I just saw someone talking about feet as the basis of Latin stress. You have to assume that feet are all bimoraic, and then you just have to remove the last syllable as extrametrical.... and then sometimes you also have to remove the FIRST syllable as inexplicably extrametrical as well. Which perfectly predicts Classical Latin stress, which now always falls on the first mora of the final syllable of the word. Except of course that if feet have two morae and you can always delete one mora as extrametrical, you can always make the last foot fall whereever you need it to fall...


[fwiw, some people apparently think the Latin stress rule is an example of what you were looking for, with weight-sensitive primary stress and rhythmic secondary stress (counted backward in two from the primary), in a way that would in practice (but presumably not in theory) lead to both iambic and trochaic patterns. But I'm not sure how robustly we can really reconstruct that]
User avatar
Nel Fie
cuneiform
cuneiform
Posts: 141
Joined: 23 May 2022 15:18

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Nel Fie »

While we're on the subject of stress rhythms: if a language has rhythmical stress, and its unstressed syllables get reduced in a way that creates undesirable homophones - would it be plausible for it to develop some form of contextual elision/epenthesis, or a meaningless affix that serves no other purpose but to shift syllables into or out of stress, and thus distinguish otherwise homophonous words? Or is something else more likely to happen?

If it's not entirely clear, here's an Examplish illustration:
Spoiler:
"tsi-" : to walk
"-ne" : 1PS=SG
"-ni" : 2PS=SG

Assuming a trochaic pattern, the vowels of the unstressed suffixes would get reduced to schwa as such, leading to both sounding the same:

tsi + ne = tsi'nə : "I walk"
tsi + ni = tsi'nə : "You walk"

and the problem could be solved as such:

"e-" : [meaningless affix]

tsi + ne = tsi'nə : "I walk"
e + tsi + ni = e'tsəni' : "You walk"

(or, alternatively, an infix could develop to preserve both verb and agreement, e.g. "-ʔe-" => tsi + ʔe + ni = tsi'ʔəni' : "You walk".)
:deu: Native (Swabian) | :fra: Native (Belgian) | :eng: Fluent | :rus: Beginner
DeviantArt | YouTube | Tumblr
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Creyeditor »

The affix would probably not be meaningless. It would probably derive from a meaningful morpheme (maybe a pronoun) and it would gain meaning in its context, because it would distinguish two forns that would otherwise look identical. Apart from that I believe it is possible.
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
Nel Fie
cuneiform
cuneiform
Posts: 141
Joined: 23 May 2022 15:18

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Nel Fie »

Good to know, thank you very much!
:deu: Native (Swabian) | :fra: Native (Belgian) | :eng: Fluent | :rus: Beginner
DeviantArt | YouTube | Tumblr
Salmoneus
MVP
MVP
Posts: 3030
Joined: 19 Sep 2011 19:37

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Salmoneus »

Nel Fie wrote: 31 Jul 2022 14:32 While we're on the subject of stress rhythms: if a language has rhythmical stress, and its unstressed syllables get reduced in a way that creates undesirable homophones - would it be plausible for it to develop some form of contextual elision/epenthesis, or a meaningless affix that serves no other purpose but to shift syllables into or out of stress, and thus distinguish otherwise homophonous words? Or is something else more likely to happen?

If it's not entirely clear, here's an Examplish illustration:
Spoiler:
"tsi-" : to walk
"-ne" : 1PS=SG
"-ni" : 2PS=SG

Assuming a trochaic pattern, the vowels of the unstressed suffixes would get reduced to schwa as such, leading to both sounding the same:

tsi + ne = tsi'nə : "I walk"
tsi + ni = tsi'nə : "You walk"

and the problem could be solved as such:

"e-" : [meaningless affix]

tsi + ne = tsi'nə : "I walk"
e + tsi + ni = e'tsəni' : "You walk"

(or, alternatively, an infix could develop to preserve both verb and agreement, e.g. "-ʔe-" => tsi + ʔe + ni = tsi'ʔəni' : "You walk".)
The first thing I'd note is that the example isn't about homophony per se (words merging), but about paradigmatic syncresis (the same inflection being able to mark more than one thing). Homophony might be resolved on an ad hoc basis, word-by-word; but syncretic affixes are going to be resolved systematically, or not at all.

Usually, it will be 'not at all'. Syncresis - even complete loss of morphological categories - happens all the time as the result of sound changes (and sometimes even without the sound changes). C.f. the loss of case in English.

In your example, there's really very little being lost: context will, 99% of the time, make clear whether I am walking or you are walking, because the listener will know whether they are walking or not. For instance, "walk" in English does not distinguish between 1st singular and 2nd singular.

Having said that, it's true that, sometimes, if people feel that an important distinction is being lost, they will innovate ways to maintain it. Importantly, they will not simply "add a meaningless particle to move the stress" - people don't think so analytically about casual word use. However, if an alternative construction DOES exist in their language, and that construction does happen to avoid the 'problematic' merger, it may become more popular than the original construction. In this case, a prefix that serves to mark a second-person subject could be added optionally (perhaps it means "sir" or "you over there" or the like, or "I hear that" (to avoid impoliteness), or even just an adverb like "forth" or so on), and the prefixing constuction could come to be prefered in part because it retains some information that the non-prefixing construction loses.

Then again, having also said that, I'd just point out that in practice your solution creates a much bigger problem: it prevents confusion between "you walk" and "I walk" (almost never confused anyway!) but creates confusion between "you walk" and "you fornicate" and "you dribble", or whatever *tsu and *tsa mean. That is, if you're going to be reducing a vowel to schwa in a way that causes morphemic mergers, you'd probably prefer to be merging inflectional affixes (which have low semantic load) than to be merging lexical items (which have very high semantic load).


Imagine if this happened in English:
- scenario A, "pat" can mean either "I pat" or "you pat"
- scenario B, "youpt" can mean "you pat", "you pout", "you put", "you putt", "you pot", "you pit", "you pet", "you part" or "you peat"

Which of these would be more confusing?
User avatar
Nel Fie
cuneiform
cuneiform
Posts: 141
Joined: 23 May 2022 15:18

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Nel Fie »

Thank you for taking the time for such an in-depth comment on the matter!

I have to apologize as well since you took such a thorough look at the example I gave. It was not intended as an actual implementation, only as a rough illustration, in case my description of the idea wasn't clear enough.
So, as far as I'm concerned, the problem of verb stems getting mixed up could (in theory) be solved quite easily by implementing the concept as an interfix rather than as a prefix, e.g. "-ʔe-" => tsi + ʔe + ni = tsi'ʔəni' : "You walk".

That said, my concern here is not the theoretical implementation and the possible knock-on effects (though it would certainly be a fascinating subject to explore), but the naturalistic plausibility of it.

So, to conclude, I can write this one down as:

"It's unlikely to happen, and if it does, it must be grammaticalized* from an existing lexical item - but it's more likely be solved pragmatically (e.g. context), or by clearing up the intended meaning with an extended construction (morphological, phrasal, sentential or otherwise) which will not necessarily be grammaticalized to serve that sole purpose."

Correct?

* Using "grammaticalized" in the broadest sense of "being divested of its original purpose".
:deu: Native (Swabian) | :fra: Native (Belgian) | :eng: Fluent | :rus: Beginner
DeviantArt | YouTube | Tumblr
teotlxixtli
cuneiform
cuneiform
Posts: 85
Joined: 05 Jan 2021 04:37

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by teotlxixtli »

What kind of linguistic features always co-occur? Which ones are mutually exclusive? Although I’m sure it exists I can’t think of a language off the top of my head that has both tone and grammatical case (for example)
User avatar
Nel Fie
cuneiform
cuneiform
Posts: 141
Joined: 23 May 2022 15:18

Re: (Conlangs) Q&A Thread - Quick questions go here

Post by Nel Fie »

Though I'll defer to anyone with more experience, I think that a list of co-occurences can either be very long because of the sheer number of possible feature combinations; or extremely short because there aren't many "always" in linguistics, as far as I can tell. There are strong tendencies for certain features to occur together, but that doesn't mean that they have to, or that they always will.
Same with the reverse - and your wording suggests at least some kind of answer here: some features are mutually exclusive on a logical basis, e.g. you cannot simultaneously "have a case system" and "not have a case system". But beyond that there seems to be plenty of room even for unlikely combination.

Picking up on your own example, WALS provides Oromo and Koasati as languages with morphological case marking and complex tone systems. You'll find more if you look at further combinations of the features "49A: Number of Cases" and "13A: Tone". (linked above on the word "WALS")

There seems to be indeed a lesser number of languages with both tone and morphological case marking going by WALS's numbers, but that doesn't necessarily imply a correlation, since by far the most common are languages that have neither, both independently and in combination. There might be some more minor trends that could be derived by crunching the numbers, but I'm currently out of time for that.
:deu: Native (Swabian) | :fra: Native (Belgian) | :eng: Fluent | :rus: Beginner
DeviantArt | YouTube | Tumblr
Post Reply