[Apertium-turkic] cases in tatar language (and in turkic languages)

[dinar q. <qd...@gm...>]

i have made a table of features of affixes and postpositions:
https://docs.google.com/spreadsheets/d/1rhSWtvLjcbCsTOLoPogEUc4SyCK4XQr_pFtqJl0eftQ/edit
.

i have sorted this things (affixes, affix+postposition, affix
combinations, all they appending to noun-likes and producing
noun-likes ) 2 ways: 1) by their level of similarity either to
separate word or to affix. 2) whether they produce semantically adverb
or adjective, or both or neither. 3) by their usual position in chain
from nominal to adverbial.

1st way, from separate words, shown with their suffixes, which are
used with them, to suffixes:

1. separate words, actively used:
га бәйле, дан башлап, га//ның чаклы, га карамастан, га караганда, /ның
буенча, /ның хакында, /ның хакындагы, ка каршы, тан башка, тан элек,
га/ның тикле, га/ның кадәрле, /ның асты, /ның өсте, /ның алды, /ның
арты, /ның яны, чагы, чак, /ның арасы, /ның буе, /ның уртасы, /ның
тышы etc

2. separate words, form changed (shortened):
дан бирле ( <- биреле), /ның турында ( <- турысында), /ның шикелле
(шәкелендәге/*шәкеленле)

3. having similar separate word with different meaning and/or used
with different suffixes:
/ның аша (ны ашып), дан алып (ны алып), га бәйләнешле (бау белән
бәйләнешле әйбер - кулланылмый), /ның төсле (... төстәге), /ның буйлап
(... буеннан/*буенлап), /ның хакта (/ның хакында), /ның аркасында
(/ның тәненең аркасында - башка мәгънә), дан соң (ның соңында), дан
тыш (ның тышында), /ның аркылы (га арткылы)

4. requiring additional suffix, etymological source word is visible
and having their own stress like separate words:
га таба (таб-а), га хәтле (хәттин ашу)

5. requiring additional suffix and having its own stress, like separate words:
га кадәр

6. requiring additional suffix after pronouns, etymological source
word is visible and having their own stress, like separate words:
/ның саен (сан), /ның хәтле (хәттин ашу), /ның сыман (аксым)

7. changing depending on previous word, like suffixes in tatar,
but requiring additional suffix after pronouns, etymological source
word is visible and having their own stress, like separate words:
/ның ча (чак)

8. requiring additional suffix after pronouns, etymological source
word is visible and having their own stress, like separate words:
/ның өчен (чөнки)

9. requiring additional suffix after pronouns and having their own
stress, like separate words:
/ның белән, /ның кадәр

10. changing depending on previous word, like suffixes in tatar,
but etymological source word is visible, having their own stress, like
separate words, and usually it is only 1 time in a sentence:
да (тагын)

11. changing depending on previous word, like suffixes in tatar, and
also do not have their separate stress,
but do not pull stress to themselves, and usually every of them is
only 1 time in a sentence:
генә, үк, мы (мы can be 2 times... : иртәме-соңмы, тегеләйме-болаймы)

12. real suffixes, (but i have included here several suffix combinations):
га, тан, та, дай, дый, ның, ны, сыз, лы, чы, ла-й, лы-й, ла-п, ныкы,
кы, рак, сыл, гылт, чыл, чән, лек, да-гы, лә-р, лә-ячәк, л-и-се, лә-ү,
лә-ү-че, лар, ы, м, ң, ыбыз, ыгыз, лар-ы

then, since there are too many in the 12th group, i have grouped them
into groups by their semantic: adverb or adjective, or both or
neither, trying to find from them "cases", then i i added to the 4
categories also other 11 groups:

neither adjective nor adverb:
(1) /ның асты, /ның өсте, /ның алды, /ның арты, /ның яны, чагы, чак,
/ның арасы, /ның буе, /ның уртасы, /ның тышы
(12) лар, ы, м, ң, ыбыз, ыгыз, лары, лә-ү

adjective and not adverb:
(1) /ның хакындагы
(3) га бәйләнешле
(12) лә-ү-че, ныкы, кы, сыл, гылт, чыл, чән, лек, да-гы, лә-р,
лә-ячәк, л-и-се, чы, ның

adverb and not adjective:
(1) дан башлап, га//ның чаклы, га карамастан, га каршы, тан башка, тан
элек, га тикле, га кадәрле
(3) /ның аша, дан алып, /ның буйлап, /ның аркасында, /ның аркылы
(4) га таба, га хәтле
(5) га кадәр
(10) да
(12) ла-й, лы-й, ла-п, га, тан, та, ны,

which can be both adjective and adverb:
(1) /ның кадәрле, ның тикле, га бәйле, га караганда, /ның буенча, /ның хакында
(2) дан бирле, /ның турында, /ның шикелле
(3) /ның төсле, /ның хакта, дан соң, дан тыш
(6) /ның саен, /ның хәтле, /ның сыман
(7) /ның ча
(8) /ның өчен
(9) /ның белән, /ның кадәр
(11) генә, үк, мы
(12) сыз, лы, дай, рак, дый

you can see that 4 of 5 case suffixes of modern tatar
academic/official grammar are in the "adverb and not adjective"
category, and they are only ones there real simple suffixes, others
are combinations with -la- . 1 left case suffix is among many other
suffixes like it.
so there is sense to remove the left "ның" from cases,
and , another way : sense to add all or many other suffixes like
"ның". in latter way, the first way's 4 suffixes are included like
category, so then it is needed just include all 4 categories, then
they are just all suffixes, not cases, so the 2nd way is excluded.
what to do with nominative case? if i remove it: then there are 4
cases by 1st way - that is just a little group of all suffixes, so it
is excluded, if i add a nominative case with "zero" suffix, i need to
add only adjective usage of nominative case, like in "китап уку", and
then question appear: if i add so-called "zero" suffix here, why not
to add also postpositions here? so also it is a nonsense and excluded.
so there are no cases.

(either term "affix" or "suffix" are used here for same thing).

there is another way to sort this things: by their position: га, then
үк, then мы ... that is already shown in grammar books and i also
should thing about that and sort that way and write blog post about
that. i think that all that things could be positioned , if they had
appropriate meaning, in any order, but get some stable positions for
semantic reasons and so probably do not show cases... it is like
arrangement of "big red book" - it could be "red big book" but it is
not used so, because there is rule that more integral/inalienable
feature's word is located more close to the main word. that rule works
also with the suffixes and postpositions: "my books " is not "{my
book} s" but "my {book s}" so also in tatar "lar" is located earlier
than "m" : "kitap lar m". in that way some suffixes get some position.
some are located even after postpositions (while postpositions are
like separate words): "gina", "uk", "da", but they can be located also
before some of separate words: "шуңа гына карап", "шуңа карап кына"
("looking only at it" and "only looking at it"). i think that
positions are not very important, and that sorting gives not only some
position of modern grammars' 5 "cases" (without nominative), but also
other positions of some other suffixes are visible. i have made an
incomplete position scheme:

[...->noun] [сыз/лы/ла-у/лә-ү-че/ныкы/рак/лык/дай ->noun] [lar]
[м/ң/ы/ыбыз/ыгыз/лары] [[ның]өсте/[ның]асты/.../дай/ныкы->noun]
[га/дан/[ның]өчен/[ның]саен/ны/ның/[ның]ча->noun/та/[ның]белән/[ның/ка]кадәр[ле->noun]/[ның]хакында[гы->noun]/габәйләнешле
-> noun] [лык/рак/булган/дай -> noun ] [ук/гына] [булган->noun] [дә]
[мы/ич/бит] [инде]

so i can say from here that 5 case suffixes are in one place: before
or in place of postposition, after posessives. "day" suffix can be in
place of them but it can be also in other places, and also nominative
case "zero" suffix is also can be in other places... so i think either
nominative case should be removed, or "day" should be added to modern
grammar's cases.

and, if to do not account being like a separate word important, all
words (also "cha", that is written jointly) in role of postposition,
except which create new "nominative/noun" and so start same semantical
order of suffixes again, like "га бәйләнешле", "(ның) өсте" and cannot
be used to describe verb, can be added to case. so, all that "cases"
are preparation do describe verb. so this explains why "шуңа гына
карап" and "шуңа карап кына" are both possible - because "карап" is
like postposition, but it is itself a verb. so, probably all words in
postposition role that are normal verbs also should not be added to
cases. "караганда" is also verb, but it has changed meaning,
lexicalised, and also it cannot be used this way: "китапка гына
караганда" - if this is used, it "immediately" changes meaning to
"кара"'s main meaning, as normal verb, to see: "when to look only at
book". "китапка караганда гына" can be used in 2 meanings: "only while
compared with book", and "only while looking at book", so, i am saying
about that 1st of these (ie "китапка караганда гына") can be accounted
as postposition usage. i feel "карап" also changes its meaning: "шуңа
карап кына" - "only regarding to that" or "only looking at that";
"шуңа гына карап" - "looking only at that", "only regarding to that" -
so, no, both meanings (looking and regarding) are here anyway... and,
"only to look at" and "to look only at" have different meanings. so,
in other hand, i should not totally exclude that verbs. i feel this is
more correct, so, if they are located after "гына" they are verbs, if
before - i am not sure: can "карап" as "looking" be accounted as
postposition? i think only "regarding" meaning should be included.

since nominative case can be used as direct object (so, describe
verb), that role can be added to cases...

i get pospositions to add from above "adverb and not adjective" and
"which can be both adjective and adverb" lists:
дан башлап - i do not add this because it can be used in "дан гына
башлап" while i am going to take all things that are before "гына"...
no... let add it..
га//ның чаклы - ok, and , "лы" is sometimes in noun role.
га карамастан, га каршы, тан башка, тан элек, га тикле, га кадәрле - ok
/ның аша, дан алып, /ның буйлап, /ның аркасында, /ның аркылы -ok
га таба, га хәтле - ok
га кадәр - ok
да - i do not include this, it is not case by definition... (it is in
later position, it is "кисәкчә")
ла-й, лы-й, ла-п - no, these are not after posessive.
га, тан, та, ны - ok. ah, there are only 4 of modern grammars' cases
and "ның" is not here in the list! so "ның" should also be removed!
because it does not describe verb! in other hand, it is located in the
same position with cases, but is used not directly do describe verb,
but only to connect pronouns to postpositions! hm! i think there
should be suffix & postposition combinations in the list , so single
"ның" will not be there.
ның кадәрле ның тикле, га бәйле, га караганда, /ның буенча, /ның хакында - ok
дан бирле, /ның турында, /ның шикелле - ok
/ның төсле, /ның хакта, дан соң, дан тыш - ok
/ның саен, /ның хәтле, /ның сыман - ok
/ның ча - ok
/ның өчен - ok
/ның белән, /ның кадәр - ok
генә, үк, мы - no (by definition ...)
сыз - no, it is not used after posessives, though it could be useful.
- i could add this as direct object of бул, и, ди but, not being after
posessive is enough to say 'no' (and others in this list are like
that). - this is mistake: it can be accounted as able to be used after
posessives: капка төбе -> капка төпсез. but it is not used with other
posessives: no китабымсыз, no китабыңсыз. but it can be: синең
китапсыз, минем китапсыз. but it cannot be: аның китапсыз, шул
капканың төпсез ( - instead of them: аның китабыннан башка/гайре/азат,
шул капканың төбеннән башка/гайре/азат). i do not include it.
лы - no, it is not used to describe verb, белән is used instead, and
it is not used after posessives, though it could be useful. - this is
mistake: it can be accounted as able to be used after posessives:
капка төбе -> капка төпле. but it is not used with other posessives:
no китабымлы, no китабыңлы. also no синең китаплы, no минем китаплы.
(instead of them: китабыма ия, китабыңа ия, синең китабыңа ия, синең
китабыңны алган/тоткан, минем китабыма ия, минем китабымны
алган/тоткан). i do not include it.
дай - ok
рак, дый - no, they are not used after posessives. (гә охшаганрак is
used instead of рак after posessive: not күктөсерәк, but күк төсенә
охшаганрак).

if i am going to include direct objects roles, i should look to
another/left 2 categories, and i see neither of postpositions in them
are used in that role, except "чы", which is also really not used so.
- i made mistake here. some of them are direct objects for бул
(become, be), и (be), ди (say). i should recheck them.

and if i add direct objects, i should also add subject, because i
think it just describes verb. argument/proof : "ул агач су буенда үсә"
- subject is at first place here, but in "су буенда агач үсә" it is
more close to verb and it is just describes it as adverbials, [i
think] it is not just poetic order of words, it is just normal order
of words. but if i add subject, i have a problem: so it descibes verb,
so i should find and add also suffixes/postpositions which create
forms which are used as subjects.

so i look the first 2 categories again for subject roles - there was
mistake, i will check also for direct object of бул, и, ди:
/ның асты, /ның өсте, /ның алды, /ның арты, /ның яны - no, because
they already have "ы" posessive suffix
чагы, чак - no, not used after posessives, usually
/ның арасы, /ның буе, /ның уртасы, /ның тышы - no
лар, ы, м, ң, ыбыз, ыгыз, лары, лә-ү - no, posessive ones by
definition; лә-ү, лар are not used after posessives. (additionally лар
is usually not used with бул, и, ди);
/ның хакындагы - no, not used as subject (requires a noun after
itself) - that was mistake; but it can be direct object of бул, и, ди,
so, ok.
га бәйләнешле - no, not used as subject - that was mistake; but it can
be direct object of бул, и, ди, so, ok.
лә-ү-че - no (not used after posessives)
ныкы -ok - subject, also can be direct object of бул, и, ди;
кы, сыл, гылт, чыл, чән, лек - no (өчен, өчен булган is used instead
of лек in that position) - they can be direct object of бул, и, ди but
they are not used after posessives.
да-гы - no, it is used after posessive in form да-гы-сы - ie with
additional posessive suffix after it - that was mistake; but it can be
direct object of бул, и, ди, so, ok.
лә-р, лә-ячәк, л-и-се, чы - no (not used after posessives)
ның - no (not used as subject) (but it can be included in cases by
some definitions of cases... ) (also not direct object of бул, и, ди)

and i search subjects from suffixes and postpositions marked with "no"
from last 2 categories - there was mistake, i will check also for
direct object of бул, и, ди - no, i have already checked them.
ла-й, лы-й, ла-п - no, these are not subjects. (- they can be direct
object of бул, и, ди, so, ok (эшли иде, эшләп булган), but , no, they
cannot be after posessive).
генә, үк, мы - no (by definition ...)
сыз - no, it is not used after posessives. and it is usually not used
at subject position.
лы - no, it is not used after posessives. rarely used at subject position.
рак, дый - no, they are not used after posessives.
(дай is already 'ok'. (but in this check (for subject) it would be
'no', it is not usually used at subject position). (it can be direct
object of бул, и, ди).)

so, only ныкы is added, after checking for subject role. but there is
a strange thing with ныкы: китапларыныкы йөри - ok, it is subject.
китапларыныкы белән йөри - it made a noun and the chain can be
restarted from it: китапларыныкылар белән йөри. strangeness becomes
visible if i say китабымныкы. китабымныкы йөри - ok, it is subject.
китапларыныкы белән йөри - it made a noun and белән is added at new
chain, ok. but strangeness becomes visible if i use "га кадәр":
китабымныкыга кадәр - i feel it strange and i use китабымныкыма кадәр.
other posessives: китабыбызныкыга кадәр, китабыбызныкыбызга кадәр -
this time it works normally. ok, i will include ныкы. and anyway (a
fortiori) i do not add its this usage, but only when there are no
other cases after it and it is used as subject...

so, i add several things after checking for being direct object of бул, и, ди:
[ның] хакындагы, га бәйләнешле, да-гы. ([ның] хакында is already included).

so i make several definitions and lists of cases:
1) suffixes that can be located after optional posessives and before
optional postpositions: ның, ны, да, га, дан, дай, ныкы, дагы - 8
items.
1a) suffixes that can be located only {after optional posessives and
before optional postpositions}: ның, ны, да, га, дан - 5 items.
2) word forms that can be used after optional posessive suffixes and
before optional postpositions: nominative (no suffix), ның (form
marked with ның), ны, да, га, дан, дай, ныкы, дагы - 9 items.
2a) word forms that can be used only {after optional posessive
suffixes and before optional postpositions}: ның (form marked with
ның), ны, да, га, дан - 5 items. (like in 2nd definition).
3) verb describer/ing roles that can be used after optional posessive
suffixes and before optional postpositions: indefinite direct object
(no suffix), subject (no suffix), ны (role marked with ны), да, га,
дан, дай, ныкы, дагы - 9 items. (ның is not here because it is not a
role, but just a helper suffix that must be used after pronouns before
prepositions).
3a) verb describer/ing roles that can be used only {after optional
posessive suffixes and before optional postpositions}: ны (role marked
with ны), да, га, дан - 4 items. (ның is not here because it is not a
role, but just a helper suffix that must be used after pronouns before
prepositions).
4) combinations of suffixes with postpositions located after optional
posessives and before optional "кисәкчә" (ук/гына/дә/мы/...), with
exception of combinations with postpositions which are normally used
verbs (they can be used also after "кисәкчә")... verbs with changed
meaning can be included, even if the changed meaning can also be used
after "кисәкчә" (example: "карап"):
га, тан, та, ны, дай, ныкы, дагы, чаклы, га чаклы, ның чаклы, га
карамастан, га каршы, тан башка, тан элек, га тикле, ның тикле, тикле,
га кадәрле, кадәрле, ның кадәрле, аша, ның аша, дан алып, буйлап, ның
буйлап, аркасында, ның аркасында, аркылы, ның аркылы, га таба, га
хәтле, хәтле, ның хәтле, га кадәр, кадәр, ның кадәр, тикле, ның тикле,
га бәйле, га караганда, буенча, ның буенча, хакында, ның хакында, дан
бирле, турында, ның турында, шикелле, ның шикелле, төсле, ның төсле,
хакта, ның хакта, дан соң, дан тыш, саен, ның саен, сыман, ның сыман,
ча, ның ча, га ча, өчен, ның өчен, белән, ның белән, га карап (only
with meaning of "regarding"), дан башлап (this looks like postposition
compared to "ны башлап" , "белән башлап"), ның хакындагы, хакындагы,
га бәйләнешле - 71 items but probably this list is not complete.
4a) 4 excluding дай, ныкы, дагы.
5) verb describer roles used after optional posessives and before
optional "кисәкчә" (ук/гына/дә/мы/...), with exception of roles made
with postpositions which are normally used verbs (they can be used
also after "кисәкчә")... postpositions consisting of verbs with
changed meaning can be included, even if the changed meaning can also
be used after "кисәкчә" (example: "карап"):
indefinite direct object (no suffix), subject (no suffix), га (role
marked with га), тан, та, ны, дай, ныкы, [ның] чаклы, га чаклы, га
карамастан, га каршы, тан башка, тан элек, га тикле, [ның] тикле, га
кадәрле, [ның] кадәрле, [ның] аша, дан алып, буйлап, [ның] аркасында,
[ның] аркылы, га таба, га хәтле, [ның] хәтле, га кадәр, [ның] кадәр,
[ның] тикле, га бәйле, га караганда, [ның] буенча, [ның] хакында, дан
бирле, [ның] турында, [ның] шикелле, [ның] төсле, [ның] хакта, дан
соң, дан тыш, [ның] саен, [ның] сыман, [ның] ча, га ча, [ның] өчен,
[ның] белән, га карап (only with meaning of "regarding"), дан башлап
(this looks like postposition compared to "ны башлап" , "[ның] белән
башлап"), [ның] хакындагы, га бәйләнешле - 50 items but probably this
list is not complete.
5a) 5 excluding indefinite direct object, subject, дай, ныкы, дагы.
6) to write postpositions jointly, and resulting "words" could be
accounted "word forms", and here is definition for cases then:
word forms made with combinations of suffixes with postpositions, used
after optional posessives and before optional "кисәкчә"
(ук/гына/дә/мы/...), with exception of forms made with postpositions
which are normally used verbs (they can be used also after
"кисәкчә")... verbs with changed meaning can be included, even if the
changed meaning can also be used after "кисәкчә" (example: "карап"):
nominative(no suffix), га (word form made with га), тан, та, ны, дай,
ныкы, [ның] чаклы, га чаклы, ........
6a) 6 excluding nominative, дай, ныкы, дагы.

("...дай" means "[that/which is] like ...", "...ныкы" means
"something/that/which is of ...", "...дагы" means "[that/which is] at
...").

(square brackets mark optional things).

i like definitions 3 and 5.

by all these definitions, cases are verb describers / adverbials. so
"ның" is not among them. but in european languages and in arabic there
is case with meaning of "ның". and i should include it somewhere. and
it is in same position, but it is noun-desciber, it is like adjective.
so i can include them by editing above definitions or by adding
additional definitions. i do not edit them. even that ның position
should be accounted as another thing, because it is not in
noun-to-adverbial chain/order, and there cannot be posposition after
it, and teher is noun with posessive after it, not verb.

as i said, there are also other positions of suffixes are visible. i
am going to distinguish them.

since there are "ның" in arabic cases and in russian cases, maybe,
other positions/roles also could be included into cases in tatar
language? - if to include, they will are another type of cases,
additional to the above defined type of case, they are used before or
after using the above defined cases.

i account булган and икән as words made from verbs бул and и. there
are other such words: булды, булыр, булачак, иде, име.

үк, гына - "кисәкчә1".
дә - "кисәкчә2".
мы - "кисәкчә3".
ни, икән - "кисәкчә4". - икән here is different from verb икән: "укучы
икән", "укучы икәнмен", etc - verb, "{{китап мы икән инде} {анда}
бул}ган" - here it is "кисәкчә4", i think.
инде - "кисәкчә5".
these can be latest levels of describers of verb, (after
postpositions), but they also can be used after verb with its suffixes
("булган да мы инде", "укучы икәнмен дә", "укучы икәнмен инде").

levels before "case"s:

words like "өсте" - i just remove them from the "chain". they are
normally used nouns, with posessive. they produce nouns, counted and
posessed, and chain is started from beginning after that... : "өстәл
асларыбызга", "өстәл аслары кырыенда", "өстәл асларына", "өстәл
өстенекеләр".

м/ң/ы/ыбыз/ыгыз/лары - posessives.

лар

рак, лык - these are just noun-to-noun producers, but they are more
usual at this place

лы, сыз - these are just noun-to-noun producers, but they are more
usual at this place

чы - these are just noun-to-noun producers, but they are more usual at
this place

ла-у - i remove this, accounting it usage of intermediate helper verb...

there are lot of things , noun producers can stay here, i can account
them and simple nouns as first item in chain:
сыз, лы, ныкы, рак, лык, дай, [ның]ча, [ның/ка]кадәрле,
[ның]хакындагы, габәйләнешле, ( ы (posessive) could be accounted but ы
is removed/hidden if suffixes added to it, and i do not include it
here: су буе - субуйлы, субуйлап; капка төбе - капкатөпсез,
капкатөпле; - though i would like if it could be used as су буенлы, су
буенсыз, капка төбенле, капка төбенсез, and i write so sometimes by
myself ). there are also verb-to-noun producers: у, ган, мак .

what to do with ның before noun with posessive, as i do not include it
into cases - it is always with posessives (except ныкы which is
probably derived from ның-кы), so it is part of their system. (usage
of ның without posessive - бу - минем китап, instead of бу - минем
китабым - is just incorrect modern usage, appeared by russian language
influence, as i know). (and also ның is not among noun-to-noun
producers, it is not used at that position, instead, ныкы is used).

see also threads in this list
http://sourceforge.net/p/apertium/mailman/message/29485329/ and
http://sourceforge.net/p/apertium/mailman/message/32130070 .