Having a problem with case endings...

[xBlackWolfx]

I havent posted here in a while, bc of all the trolls, but i decided to bc the forums over on unilang havent helped me much... maybe i'll get some help maybe not.

I'm not working on my pseudo-auxlang anymore just so we know, now i'm working on a fairly complex artlang. I have just one problem....

The language relies on suffixes like finnish to mark cases, most of which are just replacing the final vowel ('I' in the nominative is ku, in the accusative/partitive it's ka, and in the genitive it's kene).

my phonotactics are C[l]V(C), V can be a monophthong or a diphthong.

here is my list of consonants as it currently stands (digraphs represent affricates):

p b pf bv f v m t d s z ts dz n l r ʃ ʒ tʃ dʒ k g x ɣ

the vowel system is still underconstruction, but right now it at least contains i ɛ u o ɑ for monophthongs, and aʊ aɪ eɪ ɔɪ for diphthongs. i was going to have an german-like vowel system, but i decided too many of the vowels sounded the same, this was particularly problematic since the vowels are used to determine case.

Words are all required to begin in a consonant, and end in a vowel. This is bc if a word ends in the same phoneme (or something similar) as the next word begins with, that would make the last letter of the previous word pretty much inaudible.

This is where my problem comes from, my language will have probably more than 20 cases (the design goal is partially driven by the fact that i want it to detail alot of information with little amount of extra syllables). Effectively, most cases are formed simply by replacing the final vowel (the pronouns use a different system from the nouns, for nouns the nominative ending is -a, and the pronouns do add a syllable for most of their cases), but even with my old vowel system i dont have enough to form 20+ endings. And these arent minor cases that i can remove, most are actually rather generic and multi-use, the large number comes from the abnormally high number of concepts that they represent. Here is my list as it currently is (i havent finished it, more will come):

1.agent

2.patien (also used for where other case endings dont fit, like esperanto's je and finnish's partitiivi)

3.locative

4.case ending that functions like japanese 'no' (mostly used to form the genitive)

5.dative (may also used for when something is changed into something else, based off of japanese 'ni')

6.temporal case (when something happens, only used with single words, a conjunction is used for whole sentences)

7.desination (may remove this and just use the dative)

8.origin

9.towards/in the direction of (japanese 'e')

10.away from

11.privative (used like english's preposition 'without', also used to like the english suffix -less, though in such a case it would also take an extra suffix to mark it as describing a noun rather than describing the action, this is a common distinction that someone told me once but i forgot it, i think it was adverbal and adjectival)

12.comitative/sociative

13.tool/manner/technique (based off of japanese 'de' and also used to form most adverbs, except for temporal adverbs which take the temporal case)

14. causal (because of...)

15.beneficial (for someone, may just use the dative)

16.despite (may or may not make it in)

And more will be added as i think of them.

As we see, my artlang here does not rely on coverbs like my old conlang did. Sentences will only have one verb (auxilarly verbs will be represented with prefixes, mood and aspect are represented with suffixes).

And ontop of all of that, alot of roots double as both nouns and verbs, with the only difference being if they take a noun ending or a verb ending. Tense is marked by the final vowel of a verb, so that takes up three of them right there (right now the present progressive is marked with -e, and the negative present progressive is -embe).

I could fix this simply by allowing words to end in consonants (this wouldn't add any extra syllables, so it wouldnt lengthen the language either). But relying so heavily on final consonants to determine case and tense and aspect would cause a problem if the next word began with the same consonant or a similar one. In my case, if a certain case had to end in -t, then if the next word began with t d s z ts dz ʃ ʒ tʃ dʒ, it would be inaudible (if it was followed by one of the fricatives, then it would sound like one of the affricates). This is a huge problem.

The only remotely useful post i got on unilang told me i shouldnt worry about it (but that was from some guy who claimed to be from ireland, but i couldnt understand most of what he said, and he doesnt appear to have been able to understand me, i assume it was dialectal differences). Most languages of course get along fine with consonants in their syllable codas, but in those languages there is no threat of confusion if that consonant isnt heard. Infact in english (and i believe other european languages) it's common to muffle the last consonant of a word, like in american english "and" is often simply stated as " an' ". In my artlang, if the final consonant was omited, it would be incompresensible. It would be like ramdomly omitting a preposition in an english sentence. In most cases this would make it undecipherable.

Am i really fretting about it too much? Laadan does something similar, however only a few cases have a final consonant, most of them end in vowels, and even then only one or two are a single consonant (the others are CVC).

Is there any solution to this? I'm starting to think it's futile and i should just add an extra syllable to the end just to get this to work. Either that or switch to prepositions. It is a svo language afterall (though i am really tempted to make it a VSO language).

[Testyal]

Here's an idea- Instead of ending each word with only a vowel, how about having a vowel-consonant-vowel cluster. e.g. taking a random word I've made up- 'flik' could change to 'flale' in one of the cases.

[Golahet]

What about allowing a subset of consonants ending a word, and having an epenthetic vowel?

This is what I would do in your case:

Allow only the following consonants as word-final codas: m n f s ʃ x l r

Add /ə/ to the phonology.

Whenever the word-final coda + word-initial onset cluster are difficult to pronounce or hear correctly, the epenthetic /ə/ is added to the end of the word.

/ə/ is allowed to appear only in closed syllables when word-internal, and only if it isn't the word-final syllable.

So, maybe we have the word kam. Whenever that word is followed by a word that starts with /m/, it is changed to kamë (I'm using <ë> for /ə/).

A word like bënta is allowed to exist, since the ë appear in a closed syllable. Since ë is in a closed syllable, it can't be confused with a word-final epenthetic vowel. But a word like këm is not allowed, since if it is followed by a word that forces it to become këmë, it may result in ambiguity, since both syllables are open.

The consonants m n f s ʃ x l r are chosen because they will often be easy to pronounce as codas. And f s ʃ x may assimilate in voicing to v z ʒ ɣ with the following word's word-initial onset without it resulting in confusion.

Thus, often there will be no need for the epenthetic vowel, and the word will not have to have an extra syllable.

[xBlackWolfx]

so what you're saying is, i should just add a schwa sound to the end of a word if there is a chance that it's final consonant will be inaudible bc of the initial consonent of the next word?

i dont really like that much, it means you have to think and essentially make the word agree with the next word after it, even if it isnt actually related to the word (adjective-noun agreement is pretty common, but two words agreeing with eachother simply bc they appear one after the other?). however, i admit it doesnt feel odd to me to add a schwa in between the words, but as part of its own mora, seperate from the final consonant of the previous word.

finnish does something similar, if a word ends in a vowel and the next word begins in a vowel, apparently they automatically pronounce a glottal plosive at the beginning of the next word to seperate it from the previous one. though i dont find that very odd, alot of languages use the glottal plosive at the beginning of a word (arabic and hebrew for example dont allow words to begin in a vowel, the loan words that do like arabic and israel actually begin with the glottal plosive). infact i tend to pronounce vowel-initial words with a glottal plosive (i think it makes the words sound better, and no one can hear the difference anyway, which is sorta fun). though i wonder, is it really situational or do they just pronounce a glottal plosive at the beginning of every word that in spelling begins with a vowel? from what i've seen, almost all of it's case endings end in vowels and even the nouns themselves end in vowels, so a word with a final vowel isnt exactly a rare occurence.

earlier i was considering allowing ʔ and h to appear in onsets, which i may still do.

i did contemplate allowing some consonants to appear at the end of words. m n l and r for example can pretty much only be confused with themselves, (a word ending in r will only have a problem if the next word begins with an r, which is just one out of 24 possiblities). if i only allowed one or two of those to appear at the end of a word, then i could just severly limit the number of words that begin with the allowed consonants (or outright forbid them), this would drastically reduce the chances of an unwanted consonant cluster from occuring. and of course as you pointed out, the fricatives have almost the same advantage. alot of these however do have one universal disadvantage: they could be considered to be the initial consonant of the next word if it begins with l, yes ml- and xl- and even ɣl- are allowed onsets. pretty much the only (C)[l] that is forbidden in the onset is rl- (though it can occur inside a word, if they are parts of two seperate syllables). note that i may decide to ban L from coming right after any coronal in the onset. nl- will almost certainly be banned if nothing more.

If i was to allow l and r at the end of a word, then even with just 9 vowels, the total number of endings would be tripled to 27, which may be just the number i need if not more!

[Golahet]

so what you're saying is, i should just add a schwa sound to the end of a word if there is a chance that it's final consonant will be inaudible bc of the initial consonent of the next word?

i dont really like that much, it means you have to think and essentially make the word agree with the next word after it, even if it isnt actually related to the word (adjective-noun agreement is pretty common, but two words agreeing with eachother simply bc they appear one after the other?). however, i admit it doesnt feel odd to me to add a schwa in between the words, but as part of its own mora, seperate from the final consonant of the previous word.

It's no more difficult than choosing between "a" and "an" in English. It may even be simpler! It's sandhi.

(arabic and hebrew for example dont allow words to begin in a vowel, the loan words that do like arabic and israel actually begin with the glottal plosive).

I don't know about Arabic, but "Israel" in Hebrew is /jiɬ.rɑ:.ʔe:l/. /ʔ/ also occur elsewhere in words, though as far as I know never in codas. The only exception to the rule that all words have to start with a consonant in Hebrew is /u:/ that may also start a word.

i did contemplate allowing some consonants to appear at the end of words. m n l and r for example can pretty much only be confused with themselves, (a word ending in r will only have a problem if the next word begins with an r, which is just one out of 24 possiblities). if i only allowed one or two of those to appear at the end of a word, then i could just severly limit the number of words that begin with the allowed consonants (or outright forbid them), this would drastically reduce the chances of an unwanted consonant cluster from occuring. and of course as you pointed out, the fricatives have almost the same advantage. alot of these however do have one universal disadvantage: they could be considered to be the initial consonant of the next word if it begins with l, yes ml- and xl- and even ɣl- are allowed onsets. pretty much the only (C)[l] that is forbidden in the onset is rl- (though it can occur inside a word, if they are parts of two seperate syllables). note that i may decide to ban L from coming right after any coronal in the onset. nl- will almost certainly be banned if nothing more.

You could use the epenthetic schwa to disambiguate them too, so that Vm#lV (where # is a word boundary) becomes Vmë#lV, and V#mlV remains V#mlV.

Do you (intend to) have a full-blown self-segregating system? You don't seem to have it anyway. That's why I didn't mentioned the rule above, since I thought you didn't care (or if you cared could figure it out anyway).

[xBlackWolfx]

you seem to have misunderstood what i said at first. the a/an thing in english is essentially a form of agreement, between an artical and it's noun. you're asking me to have a noun agree with the noun right after it, even if the two have nothing to do with eachother besides being in the same sentence. For example, if i said in the language 'i gave it to him', if the patien (it) ended in the same consonant as word for 'him', then it would have to alter it's ending to agree with that noun, even though the two have nothing to do with eachother and infact can technically occur in any order or even be omitted (simply saying 'i gave it' would mean the same as 'i gave it away/i gave it to someone' in english). It may be an easy way to fix my problem, but it makes no sense! Maybe i'll do it for the adjective ending (i was considering having some form of agreement anyway), but really i'm not all that crazy about every single adjective ending in the same consonant except for the ones that are for a noun that begins in l.

i'm not sure what you mean by 'self-segretating'. you mean do like lojban does and have features that make it clear where one word begins and another ends? Not really, though in most of my conlangs i've achieved this simply by having the stress be on the first syllable (it marks the beginning of each new word that way). I might do the same, though i'm thinking that the old 'second-to-last-syllable' thing everyone uses is the most appealing and easiest. for example the word for 'to know' is tsixe, and 'to not know' is tsixembe (if you care, it's root word, tsixa, means 'knowledge', and i'm considering making the past progressive -au, but i dont know if like the way that sounds). i may however end up using the word-initial stress pattern to show where each word begins.

what i could do is use vowel endings for the more common cases, and use the consonant-final cases for the less used ones, and require them to appear at the end of a sentence so their final consonant can't possibly mesh with the onset of another word. of course if i plan to do this to a less-commonly used cases, i could just as well simply add a full syllable to them since most sentences wont contain them anyway. finnish appears to work the same way, most of it's endings are vowels (though technically they're all CV and CCV) with only two or three that end in consonants (in their case, it appears to only be n, and even it's verb conjugations are limited to n and t in the coda, but that's based off of the charts i can find online, i dont actually know a word of finnish). But finnish may not have such a problem since it differentiates germinated consonants, which may not occur at the beginning of a word, and all words are stressed on the first syllable. so if a stressed syllable begins with what appears to be a germinated consonant, in reality the previous word ends in that consonant.

and i was sorta looking forward to having a free-order for poetic reasons. of course common speech would have a specific order everything should go in, but for poetic reasons you would be allowed to alter the information flow to your liking.

[Golahet]

the a/an thing in english is essentially a form of agreement, between an artical and it's noun.

No. "a(n)" agrees phonologically with whatever word follows it, which accidentally sometimes happen to be the noun, sometimes not.

i'm not sure what you mean by 'self-segretating'. you mean do like lojban does and have features that make it clear where one word begins and another ends?

The same ends, but hopefully not the same means. Having such a phonology and morphotactics that word boundaries and morpheme boundaries are unambiguous. (I don't know all the details about Lojban, though I know enough to say that Lojban is an example how you should not design a self-segregating system.) Your "all words are stressed on the initial syllable" or "all words are stressed on the penultimate syllable" are good and valid partial solutions.

what i could do is use vowel endings for the more common cases, and use the consonant-final cases for the less used ones, and require them to appear at the end of a sentence so their final consonant can't possibly mesh with the onset of another word.

This doesn't work. It would mean having only one such word in any sentence, because only one word could be the last. It also means that sentences can't use recursion involving those words either. It would put an unacceptable low higher limit on the complexity of sentences. (Though I don't know the grammar of you conlang, maybe it wouldn't be a problem.) And it still doesn't deal with clusters at sentence boundaries.

of course if i plan to do this to a less-commonly used cases, i could just as well simply add a full syllable to them since most sentences wont contain them anyway.

That's probably the best solution overall: map the shortest phonological forms to the most frequent morphemes, and then accept that those less common are a syllable longer. This is my favorite solution, which I use for most of my conlangs, together with the requirement that all words end in vowels (or diphthongs).

[clockworkbanana]

the a/an thing in english is essentially a form of agreement, between an artical and it's noun.

No. "a(n)" agrees phonologically with whatever word follows it, which accidentally sometimes happen to be the noun, sometimes not.

This. I don't know about you BlackWolf, but I say "an orange banana" (should I ever need that phrase), not "a orange banana", but the latter would be what you'd get if a/an agreed with its noun.

But if you don't take this idea, I might for some language.

[Vosel]

i was going to have an german-like vowel system, but i decided too many of the vowels sounded the same,

What do you mean?

Words are all required to begin in a consonant, and end in a vowel. This is bc if a word ends in the same phoneme (or something similar) as the next word begins with, that would make the last letter of the previous word pretty much inaudible.

No it wouldn't unless you define in your phonology it should disappear.

I could fix this simply by allowing words to end in consonants (this wouldn't add any extra syllables, so it wouldnt lengthen the language either). But relying so heavily on final consonants to determine case and tense and aspect would cause a problem if the next word began with the same consonant or a similar one. In my case, if a certain case had to end in -t, then if the next word began with t d s z ts dz ʃ ʒ tʃ dʒ, it would be inaudible (if it was followed by one of the fricatives, then it would sound like one of the affricates). This is a huge problem

This is not true, they don't neccesarily disappear, it depends on how you define the phonology. Such differences are comonly made in languages. For example czech words "věci" /vjɛt͡sɪ/ (things) and "vědci" /vjɛtt͡sɪ/ (scientists) are distinct. Across word boundaries, such distinction is almost universal among languages that allow their words end in consonants.
In czech, even t+s or t+ʃ do not change into affricates unless another consonant follows, but that is AFAIK quite uncommon among languages.

Is there any solution to this? I'm starting to think it's futile and i should just add an extra syllable to the end just to get this to work. Either that or switch to prepositions. It is a svo language afterall (though i am really tempted to make it a VSO language).

If you don't want to allow word final consonats you may use -VCV endings for some cases.

[xBlackWolfx]

What do you mean?

my conlang originnally had these monophthongs:

iː ɪ ɛ æ ʌ uː ʊ oː ɑ

and these may have been added in as well:

yː ʏ øː œ

german was the closest thing that i could compare it to.

i decided to remove them since i could barely hear the difference in some of them myself. i wanted the vowels to be more distinct from eachother, so i reduced it to five monophthongs.

This is not true, they don't neccesarily disappear, it depends on how you define the phonology. Such differences are comonly made in languages. For example czech words "věci" /vjɛt͡sɪ/ (things) and "vědci" /vjɛtt͡sɪ/ (scientists) are distinct. Across word boundaries, such distinction is almost universal among languages that allow their words end in consonants.
In czech, even t+s or t+ʃ do not change into affricates unless another consonant follows, but that is AFAIK quite uncommon among languages.

i have considered that i should just ignore this problem, and allow words to end in any consonant, like you said. which i may as well do, it's not like anyone's going to speak it anyway. i intend to keep it to myself. it'll pretty much only be seen in writing.

[Systemzwang]

finnish does something similar, if a word ends in a vowel and the next word begins in a vowel, apparently they automatically pronounce a glottal plosive at the beginning of the next word to seperate it from the previous one.

Finnish doesn't do that.

though i wonder, is it really situational or do they just pronounce a glottal plosive at the beginning of every word that in spelling begins with a vowel?

the spelling isn't important per se - if they never have a word that actually starts with a vowel, they never do, regardless whether they're spelled with initial vowels or not.

[Systemzwang]

i dont really like that much, it means you have to think and essentially make the word agree with the next word after it, even if it isnt actually related to the word (adjective-noun agreement is pretty common, but two words agreeing with eachother simply bc they appear one after the other?).

this, however, does happen in loads of languages - Finnish does it to some extent, retroflex dialects of Swedish have it occuring regularly, all kinds of sandhi etc ...

See, the way Finnish works is actually: some words end in an actual silent "glottal stop" (most words ending in -e, for instance, and pretty much all imperatives9. If you pronounce such a word in isolation, nothing appears there. If you pronounce the word before a word beginning in a consonant, the consonant is lengthened (although /h/ can be left unlengthened.) Finally, if you pronounce the word before a word beginning in a vowel, a glottal stop *may appear* there.

[CrazyEttin]

finnish does something similar, if a word ends in a vowel and the next word begins in a vowel, apparently they automatically pronounce a glottal plosive at the beginning of the next word to seperate it from the previous one.

Finnish doesn't do that.

It's not really a glottal plosive, but there's certainly a short pause between the words in that kind of situation.

[xBlackWolfx]

irregardless, that still doesnt solve my problem much.

i tried a little experiment, pronouncing the word: tarxat tuna (tarxa is a word in my language, meaning hate or anger, the ending and the word tuna are just being used for the test, they have no meaning in the language as of this time).

i hear no difference between tarxat tuna, and tarxa tuna

conclusion: i obviously need to limit or forbid consonant endings if i want to understand the language myself.

and even if i do just have some of the inflections add a syllabe, that would sorta defeat my goal of making the language more short-winded. there's only 9 vowel endings, three of which will be taken up by the three tenses, so that's six endings left and my last of noun cases already has twice that number....sigh.

however i have just thought of a way that i can increase the number of possible endings without increasing the number of syllables in the word: tone. i dont like using tone much (i dont like forcing a specific rythmn or intonation on a sentence). but if i had the position of the tone make a difference, this would double my possible endings from 9 to 18. so basically half of the inflections would let the tone of the word stay where it is (beginning of word or next-to-last-syllable), the other half could have the tone moved to the last syllable. so tarxa and tarxá would mean different things.

or better yet, i should just abadon my goal here and just use prepositions like most other languages do... then again, i could have some cases and use prepositions for others, not like that's a rare feature. or i could just break down and rely completely on adpositions of some sort.

[MrKrov]

conclusion: i obviously need to limit or forbid consonant endings if i want to understand the language myself.

Or simply don't permit them at morpheme boundaries?

and even if i do just have some of the inflections add a syllabe, that would sorta defeat my goal of making the language more short-winded. there's only 9 vowel endings, three of which will be taken up by the three tenses, so that's six endings left and my last of noun cases already has twice that number....sigh.

You are just going to have to accept you'll have some "long" morphemes somewhere if you're going to have a restricted inventory to build 'em out of.

however i have just thought of a way that i can increase the number of possible endings without increasing the number of syllables in the word: tone. i dont like using tone much (i dont like forcing a specific rythmn or intonation on a sentence). but if i had the position of the tone make a difference, this would double my possible endings from 9 to 18. so basically half of the inflections would let the tone of the word stay where it is (beginning of word or next-to-last-syllable), the other half could have the tone moved to the last syllable. so tarxa and tarxá would mean different things.

Tone can do more than double the number.

or better yet, i should just abadon my goal here and just use prepositions like most other languages do [wut?]... then again, i could have some cases and use prepositions for others, not like that's a rare feature. or i could just break down and rely completely on adpositions of some sort.

And your odd problem would be solved how exactly? Wouldn't you then just say you're having a problem with prepositions? Rinse, lather, repeat?

[reizoukin]

You could find a way to use stress. The more "important" or common cases would have set endings, whereas changes in stress placement indicate different cases or variations of the common cases. For example, /tarxa'no/ could be the "towards" case, as you mentioned, while /tar'xano/ would be the "away from" case. Or something similar.

Personally, I think it seems easier than adding in tone.

[xBlackWolfx]

and i just thought of another way to make it work: have some consonants that can only appear in syllable codas, or possibly only at the end of a word (like english's ŋ).

The only issue then is which phonemes should it be? Most likely, i'll just end up adding new ones. how about ɬ? if i do, then ɮ could also be added. I can hear a clear distinction between aɬ ʃa and a ʃa. (i could of course just use x since alot of languages that have it dont allow in an onset anyway, but i do and i like it, it's my favorite phoneme).

Or, another thing i could do is just use the already existing fricatives but make them germinted at the end of a word. I could learn to differentiate germinated fricatives.

Or better yet, just give up and use VCV suffixes. That's an easier option. I'll go test that out and see how i like it.

[Golahet]

See, the way Finnish works is actually: some words end in an actual silent "glottal stop" (most words ending in -e, for instance, and pretty much all imperatives9. If you pronounce such a word in isolation, nothing appears there. If you pronounce the word before a word beginning in a consonant, the consonant is lengthened (although /h/ can be left unlengthened.) Finally, if you pronounce the word before a word beginning in a vowel, a glottal stop *may appear* there.

Does this silent "glottal stop" contrast with the absence of any consonant? Are there any minimal pairs?

i tried a little experiment, pronouncing the word: tarxat tuna (tarxa is a word in my language, meaning hate or anger, the ending and the word tuna are just being used for the test, they have no meaning in the language as of this time).

i hear no difference between tarxat tuna, and tarxa tuna

Are you sure the final /t/ is fully realized, i.e. something like [tʰɑrxɑtʰ tʰunɑ] rather than [tʰɑrxɑt̚ tʰunɑ]?

however i have just thought of a way that i can increase the number of possible endings without increasing the number of syllables in the word: tone. i dont like using tone much (i dont like forcing a specific rythmn or intonation on a sentence). but if i had the position of the tone make a difference, this would double my possible endings from 9 to 18. so basically half of the inflections would let the tone of the word stay where it is (beginning of word or next-to-last-syllable), the other half could have the tone moved to the last syllable. so tarxa and tarxá would mean different things.

But then, if the stress is at a variable position, doesn't it fail to mark word boundaries?

Maybe you could do as Ithkuil: every syllable in a word prior to the stressed syllable has mid tone, while the stressed syllable has a contour tone that ends in a non-mid position, and the syllables that follow have that non-mid level tone. Thus the stress may fell on the ultimate, penultimate, or antepenultimate syllable at the same time as it unambiguously marks the word boundaries. (Well, this isn't exactly true. The stressed syllable in Ithkuil could also have high level tone, which breaks that system and opens it up to ambiguities, but your conlang could do it like that.)

and i just thought of another way to make it work: have some consonants that can only appear in syllable codas, or possibly only at the end of a word (like english's ŋ).

The only issue then is which phonemes should it be? Most likely, i'll just end up adding new ones. how about ɬ? if i do, then ɮ could also be added. I can hear a clear distinction between aɬ ʃa and a ʃa. (i could of course just use x since alot of languages that have it dont allow in an onset anyway, but i do and i like it, it's my favorite phoneme).

This is in my opinion an ugly solution, but since it's your conlang, it's your taste that matters.


What about having a lot of diphthongs? Every combination of two monophthongs that you have shouldn't be that difficult to pronounce.

[Tanni]

I tried to understand your problem, xBlackWolfx, but I didn't really get it. Why should a consonant get ''inaudible'' if the next word begins with the same or a similar consonant? This happens all the time in many languages. What's your mother tongue?

Having something you call ''agreement'' between subsequent words is quite common, in French, there is ''la liaison''. And celtic languages use lots of mutations.

In Turkish there are the suffixes -DA and -A, locative and dative respectively. The locative suffix is acutally -da, -de, -ta or te depending on the preceeding vowel and consonant, dative is just -a or -e, depending on the previous vowel. And there are words like seyahat (voyage), so you end up with seyahatta (locative) or seyahata (dative).

[Systemzwang]

irregardless, that still doesnt solve my problem much.

i tried a little experiment, pronouncing the word: tarxat tuna (tarxa is a word in my language, meaning hate or anger, the ending and the word tuna are just being used for the test, they have no meaning in the language as of this time).

i hear no difference between tarxat tuna, and tarxa tuna

you really need to find sound samples from a language that distinguishes them, and listen to that *until you get used to hearing the difference*. introspection in things like this is not really a valid method at all. (learning to distinguish contrasts isn't a thing that comes for free, it requires practice. something it seems as though you entirely reject)