there are no cases in turkic and finno-ugric languages

http://sourceforge.net/mailarchive/message.php?msg_id=29490187 :

( > at beginning of lines mean quoting, and phrases after them are mostly not mine)

i have posted at 2012-07-03 08:40 UTC (12:40 am MSK):

hello.
i also think there are no cases in tukic languages and probably also in most of finno-ugric languages, also in other uralic and altaic langauges.
because:

what is called case ending in indoeuropean and semitic languages is clearly divided from prepositions by that prepositions are before word, and cases are after word (case endings are at end of words, after main part of word),
and second, less clear division is by that
prepositions are not modified for different words, while cases look differently for different words, (this second rule has little exceptions, for example, english 's case suffix, it alway is same "s", and russian "о" preposition may be different for different words: "о", "об", "обо").

while in turkic languages there are no prepositions,

almost no suffixes that differ for different words, such largely as in indoeuropoean languages, for example, in russian, genitiv "suffix" may be "i", "a", "", "ey", "ogo", "ih", etc, also all other cases, while in turkic languages they differ not such strongly, but just are of little difference: "non", "nin", "don", "din", for example, for so called "genitiv".

no, the stronger difference from case endings is that
turkic case suffixes are agglutinative/clitic,
but case in indo-european are inflectional (and may be fusional), that means, main part of word of many types of nouns, always used with case ending, even in nominative case, though some class of nouns can be used with "empty" case ending, in some cases, and empty case ending can mean different case, for exampel, "stol", and "knig" in russian both has empty case endings, but "stol" is nominative case, "knig" is genitiv case of plural form.

in turkic languages, "main part" of word, (ie with "empty" ie no ending) is just a noun in nominative case, and all case suffixes are just like prepositions that are written after word instead of before, so, they are postpositions. but they differ slightly depending on word, as i said, "non", "nin", etc, same happen also with prepositions in indoeuropean languages, as i said, "o", "ob", "obo" in russian, also there are other examples: "v", "vo", "k", "ko". but they both, prepositions and postpositions, do not change word, to what they connect, but cases are not so, as i said, they do not just set near nominative case of noun, but they modify its last part (ending), so, this is why they are called cases in languages they are there truely, they can be named/called "casitive" languages, and this languages, for example, indoeuropeans, are called "inflectional" and this inflectionality is in cases. in turkic languages there are no such thing. and so called case suffixes which are written connectedly, together with noun as one word, like prefixes, and so called postpositions, which are written separately from word to which they apply, in modern turkic orthographies, they should not classified be as 2 things, but they should be classified together, as of 1 class, and all things in it are can be called suffix, postfix, posposition, all this posfixes, i will call them postfixes, differ from others slightly with different properties. for example, "cha" suffix do not get stress on itself: kita'pcha (it is not called/named/classified as case suffix in modern official grammar, but rather as suffix that creates new lexem/meaning, but in fact, its meaning is constant, so it is grammatical thing) , while "qa" gets stress on itself: "kitapqa' ", most of them get stress on itself, only several don't, one more that doesn't: "bilan": 'kita'pbilan" (it is written separately in modern orthography).
but so called postpositions and so called cases of modern turkic languages sometimes have a feature that is also in true "casitive" languages: a preposition always require a case of word, to wich it apply, that is in russian: "o knige' - "about book", where "book" must be in prepositional case, and for example, while "I" pronoun is used in english after preposition, it must be in accusative case "about me", but pronouns in english are like exception from all nouns, by this behavior, also smae feature is there in tatar language: "with me" is not "minbilan" but it is "minimbilan" ie, so called genitiv case is required, but that is exception for pronouns, like in english. also there is a postposition, maybe there also others, that require a so called case suffix to be applied to word to which it apply: "taba", which means "in direction of", require suffix of so called dative case: "maktapka taba" (maktap is school).

2012/7/3, Mikel Forcada :
> Thanks a lot, guys!
>
...
>
> I am absolutely persuaded that calling these things cases is wrong.
> There is no "nominative" case, but the absolute form of the word. And
> then the genitive, accusative,etc.. are clitic postpositions that
> attach to the last member of the NP, clearly a noun. I have argued
> about this with Basques for ages.
>
...

then i have posted, by mistake to one person's mail, at first, at 2012-07-03 15:14 UTC, then to the list at 2012-07-03 15:35 UTC:

this is another clear difference between cases and prepositions in "casitive" languages: cases are applied to each word of block of words, consisting from noun with adjectives. and i think, so, english "s" do not has this feature, so may be it is not real case... of red table in russian is krasn-ogo stol-a - case ending is at both words, while in english with "s", so called case ending, it will be "red table's" , not "red's table's" . ( http://en.wikipedia.org/wiki/Saxon_genitive )

but in turkic languages that so called case endings behave just like pospositions: they are after last word of block of words.

2012/7/3, Mikel Forcada :
...
> tr kırmızı masanın
> la rubrae mensae
> ru красного столa
> eu mahai gorriaren
...

then i have posted at 2012-07-03 18:35 UTC:

some of turkic postfixes, in meaning as i called them in a previous post, do not take stress on them and that postfixes require genitive suffix before them after pronouns: sini'ncha, sini'n bilan, sini'n artqolo, sini'n asha, sini'n ochon. all of this examples are written separately in modern orthography except "cha".
in "artqolo", "asha" there is root of word visible. "cha" might be derived from "chaq", that was assumed in a book by foat ganiyev, about tatar morphology.. "вопросы морфологии", 1980 . so, there a hypotese appear that they do not take stress to itself and behafe differently with pronouns, because they were words starting with root morphemes. but i do not recognise any root of "bilan" and "ochon", but in old tatar texts "bilan" is sometimes written "birlan", i need to look to original text, with arabic letters. there "bir" may be root morpheme "one", but why "one" is there , it is strange, "one" should be before word in turkic languages, not after. root of asha is "ash", that should require accusative case, if it would be still normal verb, not genitiv, it should be "sini asha", "kitapni asha". the verb "ash" means "to rise". postposition "asha" means through/via/across/over, "over" meaning is connected/associated with "rise".

btw, also galimjan ibrahimov, famous as tatar writer, thought as me and Mikel Forcada, that there are no cases :
http://qdb.narod.ru/kitap/fanhamtil/fanhamtil_3_2004.htm : (ahmadhadi maksudi also did not call them just "cases" as modern grammars do, but called "like cases"...):

1915 елның 30 декабрендә "Йолдыз" газеты-
ның 1580нче санында Ә.Максуди "Сарыф китап-
лары" исемле тагын бер тәнкыйть мәкаләсен бас-
тыра, бу юлы ул үз имзасын ачык куя [10].
Мәкалә Г.Ибраһимовның "Аң" журналында 1915
елгы 21/22нче санда чыккан "Татар сарыфлары"
[11] исемле язмасына җавап итеп язылган. Бу
мәкаләсендә дә Ә.Максуди 1911 елда әйткән фи-
керләрен куәтли. Терминология өлкәсендә
"шәрыкче"-традиционалист буларак, ул бөтен
Ислам дөньясында бердәм гарәпчә терминология
булырга тиеш дигән тезисны алга сөрә. Фикри
һәм фәнни кыйбласын ул Шәрыкта күрә, Г. Иб-
раһимовны исә ул Европага сукыр тәкълид кылу-
да (иярүдә) гаепли [10].
Татар тел белеме өчен элек тә, хәзер дә мөһим
булган бер мәсьәләдә - исем сүз төркеменең килеш
категориясе мәсьәләсендә дә Ә.Максуди Г.Ибра-
һимов белән бәхәскә керә. Мәгълүм булганча,
Г.Ибраһимов татар телендә килешне гомумән та-
нымый, килеш кушымчаларын аерым ярдәмлек сүз
төркеме - кушымта итеп карый [2]. Ә. Максуди
исә татар телендә "игърабка ошашлы үзгәрешләр"
барлыгын билгеләп үтә, игьраб (килеш) күрене-
шен мәктәптә өйрәнүнең әһәмиятен исбатларга
тырыша. Аныңча, татар телендә игьраб (падеж,
килеш) белән танышкан балалар килеш категория-
се "тулы хокуклы" булган гарәп һәм рус телләрен
җиңелрәк үзләштерәчәкләр [10]. Дәреслекләрне
төзү өлкәсендә Ә.Максуди мисалларны әдәбият-
тан алуга каршы чыга, чөнки, аныңча, әдәби
әсәрләрдә хаталы җөмләләр күп, һәм анда гади,
җыйнак мисал табу кыен [10].

2. Ибраһимов Г. Татар сарыфы. - Казан, 1911.
10. Максуди Ә. Сарыф китаплары [Г.Ибраһимовның "Татар сарыфлары" дигән мәкаләсенә карата]//"Йолдыз", 1915, 30 дек., №1580.
11. Ибраһимов Г. Татар сарыфлары//"Аң", 1915, №21/22, 378-390 б.

(ГАЛИМҖАН ИБРАҺИМОВНЫҢ "ТАТАР САРЫФЫ"
ҺӘМ "ТАТАР НӘХҮЕ" ХЕЗМӘТЛӘРЕ ТИРӘСЕНДӘ
ЛИНГВИСТИК БӘХӘСЛӘР, Булат Хәкимов , Фән һәм тел, №3(22) # 2004)

i know this is quite offtopic for this list...

then i have posted at 2012-07-04 22:14 UTC :

"
Here are how the two categories fit some of the criteria:
http://wiki.apertium.org/wiki/User:Firespeaker/Turkic_cases_and_pos...
"
i would write also about stresses there, though probably they are not well investigated.
and i do not completely agree with that "postpositions"' preceding form is "noun phrase with some "case" (sometimes nominative/bare) ", because, in most time, at least in tatar language, nominative/bare form is used, for now, i have in mind/remember only one, that requires dative: taba, and others, except "taba", require genitiv only from pronouns. except these 2 exceptions, always bare/nominative form is used. so i agree with that as you have written it, but should be "most times" instead of "sometimes".

i have posted at 2012-07-04 22:47 UTC :

>In fact, several of them often don't stand on their own / occur in
>nominative case. But I can't think of any that don't occur in
>accusative. E.g., үйдің үстінде отырдым "I sat on top of the house",
>but cf. үйдің үстін қарап шықтым "I examined the top of the house".
>Is one of these a postposition and one not? I think a Kazakh speaker
>would identify them as the same sort of thing (until they start trying
>to translate them to other languages).
>
>> Most of them were added simply as nouns (see the subsection at the end of
>> the "Nouns" lexicon -- I tried to move existing stems there and add new
>> similar stems to that subsection. But for sure not all are present there).
>>
>> It's clear that some kind of inflectional continuation class is needed for
>> these auxiliary nouns. But having them tagged only as nouns doesn't allow
>> the system to know that they can be used as postpositions. I haven't thought
>> about it before, because they work fine in Kazakh Tatar as they are --
>> they govern the same case (nominativ / genitiv), and doesn't matter what
>> POS-tag they have :)
>
>The "nominative" they govern is actually indefinite or non-specific
>genitive—it just looks the same as the nominative ;)

i agree, yes, "kitap" in "kitap osto" is not in true nominative case, if we accept cases. and we will have to think about second genitiv with empty postfix.
i have thought, what to do with that if we refuse cases at all, and have thought out:
"[s]i[n]" is a unique turkic postfix/posposition, only one, which have 3 arguments, while all other postfixes/postpositions have only 1 argument. 1 of these 3 arguments is always required, 2 others are optional. example:
kitab o
matematika kitab o
kamil nin kitab o
kamil nin matematika kitab o
so, this argumets are marked and positioned this way:
[X non] [Y] Z o
optional arguments, with markers which are with them, ie "non", are in squera brackets.

one more argument to support this formalisation/treating: this second genitiv, with empty postfix, appears if and only if this "[s]i[n]" postfix/postposition is used.

btw i have changed my mind, no, even i have just remembered that i thought so, i think better to call them all "postposition" rather than "postfixes". because they are applied not only to one word, but to block of words, like prepositions.
and i have written recently that almost all suffixes of turkic and finno-ugric languages should be written separately, i have written that in my personal blogs and in some forums, and also in that discussions i have said about that there are no cases in these languages.
i had written about that there are no cases in turkic and finno-ugric languages several years ago in my tatar and russian blogs, without arguments, recently i have written that arguments in that forums, and have copied them to my blog post about that as comment, i was going to write that in english, so when i have seen this topic here, i have written about it. (and now i am going to copy my arguments to a blog post).

i have posted then at 2012-07-04 22:55 UTC:

>"[s]i[n]" is a unique turkic postfix/posposition ...
it is [s](i)[(""/n)/m/ng]
i have written "o" instead of "i" in examples, because "o" sounds like
"ы" form of "i" postfix/postposition.

then i have posted at 2012-07-05 14:04 UTC (near 2 hours ago) :

"
>> A thing to note is that "noun + сыз" construction can be both attributive
>> and adverbial.
I would love to get an explanation on this. Does this mean that some
PP's (postpositional phrases) attach to nouns (attributive) and other to
verbs, adverbs and adjectives (adverbial)?
"
he probably wanted to say: may be , noun + сыз = attribute, and may be that noun + сыз = adverb. so yes. example: арусыз бара - goes without tiring. арусыз кеше - i feel it some rarely used, but it mean a man without tiring. i cannot remember now good examples when it attaches to aderbs and attributes.

"
>> Thought that it might be important for translating from / to
>> a non-turkic language.
In Euskara we have sometimes two different forms of a postposition, one
that modifies NPs and another one that modifies APs, VPs or AdvP's. is
this what you are referring to?
"
so, in euskara you have such difference, but in turkic mostly same postposition is used for them all, except verbal phrases, if you mean that it ends with verbal morphem. i think all morphems are of 2 big classes: all that like noun, and verbs. then nouns may be divided into nouns, adverbs, attributes. and there are so 4 types of postpositions, that create verb from "like noun", and vice versa, and verb from verb, and like noun from like noun.
by the way, may be there are no adverbs, attributes and nouns in turkic languages, and this is another mistake like with cases.

"
> Maybe we need a third category for
> this sort of thing, but "postposition" is the wrong word, since there
> is another category in Turkic that is more like postpositions. I
> guess what I'm arguing is that while Turkic "cases" and
> "postpositions" behave differently, there are still two categories
> that roughly correspond to what we normally thing of as "cases" and
> "postpositions": maybe we should call them "pseudo-cases" and
> "pseudo-postpositions".
Too complicated. I think we should not focus on terminology now, but
rather acknowledge the different status of these entities and work
accordingly.
"
they can be divided to more classes: also there are: additional modifiers, like "uk", "gina", that are positioned after other postpositions. also postfixes that are considered/accounted "lexical" now, can be considered as one of these classes, for example, "goch", creates soem instrument, but it is not just lexical, it is grammatical, because, it has constant menaing, or can be called very productive, active, it can apply to verb phrase, ie not to just a verb, but to verb with adverbs: for example: we can say: tiz gazo goch - thing (instrument) that helps to dig fastly. but this is not very real example.. or we can say it in way that is considered grammatical in modern grammar: tiz qaziy torgan narsa.

"
for now, i have in mind/remember only one, that requires dative: taba, and others, except "taba", require genitiv only from pronouns
"
i have remembered : qadar, chaqlo, hatli, these have very similar meaning, and can be used with both dative and nominative cases, and have different meanings, "maktap chaqlo" - big like school, "maktapka chaqlo" - until reaching school.