Standard Average Altaic

[sucaeyl]

What's up y'all? I'm attempting to create a language that would appear to be very typical of those in the Altaic sprachbund, blending the features of languages like Turkish, Mongolian and Japanese. Some trends that I've already identified are:

SOV as a primary word order
Adjective-noun
Almost exclusively suffixing
Consistent (and possibly large) case marking system
Nominative-accusative allignment
Agglutinating morphology
Dependent-marking
Vowel harmony

I feel like I'm missing a lot of detail, however, especially in the orders of morphemes within nouns and verbs. So, what do you all think are some of the most identifying traits of the so-called Altaic languages? Have I made any mistakes in the generalizations that I've already uncovered? In short, what should I do to make the most boring possible agglutinating, vowel-harmonied SOV language possible? Thanks!

[clawgrip]

relative clauses before nouns

In Japanese, adjectives are actually just verbs or verb phrases, so adjective + noun is actually relative clause + noun. Not sure if other Altaic languages do anything like this though.

Turkish seems similar, because it seems to put a past tense suffix on adjectives, but perhaps this is actually a form of the copula. I don't know.

[Imralu]

relative clauses before nouns

In Japanese, adjectives are actually just verbs or verb phrases, so adjective + noun is actually relative clause + noun. Not sure if other Altaic languages do anything like this though.

Turkish seems similar, because it seems to put a past tense suffix on adjectives, but perhaps this is actually a form of the copula. I don't know.

Yeah. It's a copula. The old form, I believe, was imek (where -mek is the infinitive suffix).

When the past tense suffix (-du, -dü, -di, -dı, -tu, -tü, -ti, -tı) is added to non-verbal stems ending with a vowel, a -y- is added before.

• kötü > kötüydü
  bad > was bad

I think this can also be written in formal writing as kötü idi. The conditional endings -sa, -se are the same, adding a -y- after non-verbal stems ending in a vowel or being written as a separate word ise.

Another thing to think about is postpositions such as Turkish için ("for"), which is preceded by the genitive case with pronouns and by the nominative by other nouns. Benim için = for me (ben-im = 1s-GEN). Babam için = for my father (baba-m = father-my).

Also, a lot of spatial relations in both Japanese and Turkish are expressed with nouns that indicate a position in relation to something. Over time these will probably weather down into postpositions and from there into case endings. The word ön means the space in front of something. Ev means "house". Evim means "my house". Evimin means "my house's" or "of my house". Evimin önü means "the space in front of my house". Evimin önünde means "in the space in front of my house" = "in front of my house". Adding the copula, we can make sentences like Evimin önündeydik = "We were in front of my house."

[clawgrip]

relative clauses before nouns

In Japanese, adjectives are actually just verbs or verb phrases, so adjective + noun is actually relative clause + noun. Not sure if other Altaic languages do anything like this though.

Turkish seems similar, because it seems to put a past tense suffix on adjectives, but perhaps this is actually a form of the copula. I don't know.

Yeah. It's a copula. The old form, I believe, was imek (where -mek is the infinitive suffix).

When the past tense suffix (-du, -dü, -di, -dı, -tu, -tü, -ti, -tı) is added to non-verbal stems ending with a vowel, a -y- is added before.

• kötü > kötüydü
  bad > was bad

I think this can also be written in formal writing as kötü idi. The conditional endings -sa, -se are the same, adding a -y- after non-verbal stems ending in a vowel or being written as a separate word ise.

Another thing to think about is postpositions such as Turkish için ("for"), which is preceded by the genitive case with pronouns and by the nominative by other nouns. Benim için = for me (ben-im = 1s-GEN). Babam için = for my father (baba-m = father-my).

Also, a lot of spatial relations in both Japanese and Turkish are expressed with nouns that indicate a position in relation to something. Over time these will probably weather down into postpositions and from there into case endings. The word ön means the space in front of something. Ev means "house". Evim means "my house". Evimin means "my house's" or "of my house". Evimin önü means "the space in front of my house". Evimin önünde means "in the space in front of my house" = "in front of my house". Adding the copula, we can make sentences like Evimin önündeydik = "We were in front of my house."

As you say, the grammar is more or less the same in Japanese except that there's no possessive suffix, and the regular copula is unlikely to indicate "we were" because there is a separate spacial locative verb.

Code: Select all

ev    in  önü   nde  ydik
ie    no  mae   ni   ita
house GEN front LOC  was

Benim için is also similar in Japanese. You say watashi no tame (ni) (1 GEN sake (DAT)) where ni is dropped if the verb is the copula.


Anyway I want to add attributive verb forms to go with the relative clause preceding the noun. I know Turkish has this, and Old Japanese had it.

[sucaeyl]

What are some ways that relative clauses are formed when they're before the head noun? I know Japanese does something like

Code: Select all

tenpura-o   tabe-ta hito
tempura-OBJ eat-PST person
"the person who ate the tempura"

What do other languages do?

Also, how would these languages order TAM and person, number, etc. within the verb? For case and number within the nouns?

[Ser]

So, what do you all think are some of the most identifying traits of the so-called Altaic languages?

You forgot an enormous dislike for word-initial rhotics. I think most words in Turkish/Korean/Japanese beginning with a rhotic are borrowings? There's also a dislike for word-initial clusters.

Also, m-t pronouns (the 1st person pronouns begin with labial consonants at least in some forms, while the 2nd person pronouns begin with dental ones, also at least in forms).

In terms of grammar, optional marking of count noun plurals in Turkish/Korean/Japanese (well, associative plurals in Japanese, whatever); in particular, numbers above 1 are followed by nouns in the ""singular"" (or simply nouns unmarked for number). Verbs in Altaic do not agree with their subjects for person and number (I don't know what the order of the morphemes that verbs do have is.).

For case and number within the nouns?

Number markers go before case markers in Turkish/Korean/Japanese.

[clawgrip]

What are some ways that relative clauses are formed when they're before the head noun? I know Japanese does something like

Code: Select all

tenpura-o   tabe-ta hito
tempura-OBJ eat-PST person
"the person who ate the tempura"

What do other languages do?

Also, how would these languages order TAM and person, number, etc. within the verb? For case and number within the nouns?

Modern Japanese just sticks the verb right before the noun and that's it, but Old Japanese had a special attributive form, e.g.

Code: Select all

Hito   ga  tenpura wo  tab-u.
person SBJ tempura OBJ eat-NONPST.CONCL
"The person eats tempura."

Tenpura wo  tab-uru  hito   nari.
tempura OBJ eat-ATTR person COP.NONPST.CONCL
"It's the person who eats tempura."

Yama     taka-shi.
mountain high-NONPST.CONCL
"The mountain is high."

Taka-ki   yama     nari.
high-ATTR mountain COP.NONPST.CONCL
"It is a high mountain."

[Legion]

I'm frankly not sure there's any structure found in all of Turkic, Mongolian, Tungunsic, Japanese and Korean that you won't also find in most of the other language families spoken in northern Asia (Uralic, Chukotko-Kamchatkan, Yenisean, Nivkh, Ainu)…

Notably, the formation of relative clauses by and large uses a gap structure (with the exception of European Uralic languages which have acquired true relative pronouns under the influence of SAE), whether they do this with explicit, specific morphology (as in Turkish), or simply with syntax (as in Japanese) is irrelevant; and futhermore, this kind of structure is the most common relative structure in the world by a large margin (more than 3/4th of all languages use that structure).

You might as well try to make a Standard Average North Asian language.



You can also note that the majority of these languages use locational comparative constructions, rather than particles (again with the exception of European Uralic languages), but again, this is what the majority of languages in the world do.

[Pole, the]

Wow, is counting Japanese and Korean as Altaic already an accepted mainstream linguistic theory? Did I sleep through it?

[Herr Dunkel]

Wow, is counting Japanese and Korean as Altaic already an accepted mainstream linguistic theory? Did I sleep through it?

Altaic itself isn't mainstream, mind you.

[Legion]

Wow, is counting Japanese and Korean as Altaic already an accepted mainstream linguistic theory? Did I sleep through it?

There are different views on what Altaic is:

—Not a thing at all.
—The macro-altaic hypothesis, which includes Turkic, Mongolian, Tungunsic, Japonic and Korean.
—The micro-altaic hypothesis, which excludes Japonic and Korean.
—James Marshall Unger proposes that Altaic is formed by Japonic, Korean and Tungunsic, excluding Turkic and Mongolian.
—It is also proposed that Japanese and Korean are more closely related to each others than to any of the other families.
—Ainu is sometimes thrown in there as well.
—A small minority proposes a larger grouping into a Ural-Altaic family.

A major obstacle to the Altaic hypothesis is the disprecency between the number of Turkic/Mongolian cognates and Mongolian/Tungunsic cognates (which is high) and that of Turkic/Tungunsic cognates (which is much lower), which points to a pattern of borrowing based on geographic proximity, rather than true cognates.

Generally, the more carefully the areal factor has been investigated, the smaller the size of the residue open to the genetic explanation has tended to become. According to many scholars it only comprises a small number of monosyllabic lexical roots, including the personal pronouns and a few other deictic and auxiliary items. For these, other possible explanations have also been proposed. Most importantly, the 'Altaic' languages do not seem to share a common basic vocabulary of the type normally present in cases of genetic relationship.

[clawgrip]

The same problem seems to exist with Japanese and Korean. It's difficult to identify many non-Sinitic cognates, despite extremely similar grammar.

[sucaeyl]

Well, I guess I am trying to make Standard Average North Asian: My primary goal is to create a language that can be readily learned by a non-linguist speaker of Japanese and English, one which bears morphosyntactic similarities to Japanese and other languages sometimes considered to be in the same sprachbund. I want to simplify and generalize the features of languages like Turkish, Japanese, Manchu, Mongolian, Korean, and even those like Ainu, Yukaghir and Nivkh. So I guess I just want make a "typical" language of North Asia that would not seem out of place if spoken by a group anywhere between Kazakhstan and the Sea of Okhotsk.

[clawgrip]

Naturally if this is supposed to be serious comparative linguistics I think this whole thing is dubious, but if it is simply part of some creative conlang endeavour, then I think it is fun and am happy to contribute.

Japanese uses yori ("from") for its locational comparative construction, e.g. Neko wa nezumi yori ookii. "Cats are big from mice." There is another word for "from" but it can't be used for comparatives.

[sucaeyl]

Naturally if this is supposed to be serious comparative linguistics I think this whole thing is dubious, but if it is simply part of some creative conlang endeavour, then I think it is fun and am happy to contribute.

Just a creative endeavor! Maybe a little more background would help to show what it is that I need assistance with. My girlfriend agreed to learn a language of my creation, and, after explaining the traditional morphological typologies (which I summarized as isolating, agglutinating and polysynthetic) she decided that she would be willing to learn "a really simple gluten-whatever language, like Japanese [her L1] or Turkish". So that's what I'm attempting to create. Although heavily simplified I'd like the features that this lang includes to resemble those of real "Altaic" languages, particularly paying attention to it having features that would naturally cluster with each other. Deciding these features is where I lack the most knowledge, and subsequently where I need the most help. But any contribution is appreciated! And thanks to those who've already pitched in!

[Ithisa]

My conlang (which I posted on CBB but not here) is pretty much a Standard Average Altaic relex with some disambiguators that allow Latin-style free-word-order to be used. It is difficult to think of a, say, Japanese sentence that cannot be grammatically translated, word by word, case by case, clitic by clitic, postposition by postposition into Miyasalemmi.

[clawgrip]

My conlang (which I posted on CBB but not here) is pretty much a Standard Average Altaic relex with some disambiguators that allow Latin-style free-word-order to be used. It is difficult to think of a, say, Japanese sentence that cannot be grammatically translated, word by word, case by case, clitic by clitic, postposition by postposition into Miyasalemmi.

How about this one?

顔の見えない、見知らぬ相手だからこそ、言葉にはより気をつけないといけない。
Kao no mienai, mishiranu aite da kara koso, kotoba ni wa yori ki o tsukenai to ikenai.
face SBJ.ATTR see-POT-NEG, unknown partner COP from is.indeed word DAT SBJ from/more feeling/disposition OBJ attach-NEG COMPL go-POT-NEG
It is precisely because you are speaking to someone you do not know and whose face you cannot see that you must be careful with your words.

This sentence is interesting because it incorporates several of the things brought up in the thread, as well as some things that may be particular to Japanese:
Of course it has SOV word order, adjective-noun, exclusively suffixing, nominative-accusative, agglutinating morphology.

It has case-marking (kind of): kao no, kotoba ni, ki o

There are two instances where relative clauses take special attributive forms:
conclusive attributive
kao ga mienai kao no mienai
can't see the face (whose) face can't be seen

mishirazu mishiranu
is unknown (that) is unknown

verb compounding:
miru + shiru + nu mishiranu
see + know/realize + NEG.ATTR be.unknown

locational comparative construction (that has been abbreviated to mean "more" and is actually the only instance I know of in Japanese where a particle does not follow anything)
ki o tsukeru yori ki o tsukeru
feeling/disposition OBJ attach from feeling/disposition OBJ attach
be careful be more careful

topic-comment structure (this topic is explicitly in the dative):
kotoba ni wa
word DAT TPC

Object promoted to subject in potential, evidence of the confusion of subject and object in Japanese:
kao o miru kao ga mieru
face OBJ see face SBJ see-POT

idiomatic expression with ki:
ki o tsukeru
feeling/disposition OBJ attach
"be careful"

periphrastic expression for necessity (can be abbreviated to just the complementizer)
-nai to (ikenai)
NEG COMPL (go-can-NEG)
"must"

[Ithisa]

My conlang (which I posted on CBB but not here) is pretty much a Standard Average Altaic relex with some disambiguators that allow Latin-style free-word-order to be used. It is difficult to think of a, say, Japanese sentence that cannot be grammatically translated, word by word, case by case, clitic by clitic, postposition by postposition into Miyasalemmi.

How about this one?

顔の見えない、見知らぬ相手だからこそ、言葉にはより気をつけないといけない。
Kao no mienai, mishiranu aite da kara koso, kotoba ni wa yori ki o tsukenai to ikenai.
face SBJ.ATTR see-POT-NEG, unknown partner COP from is.indeed word DAT SBJ from/more feeling/disposition OBJ attach-NEG COMPL go-POT-NEG
It is precisely because you are speaking to someone you do not know and whose face you cannot see that you must be careful with your words.
Nan kimos takeremánkae esti ótúremáne mirinıla tema miya, lemmisihino mimikiremánakyemán.
<Rel. clause start> face-ACC see-POT-NEG-3rdPSingObj-ATTR <conjunction> know-NEG-ATTR friend-COP from precisely, word-PLUR-DAT-TOP-care-HAVE-NEG-POT-NEG.

This sentence is interesting because it incorporates several of the things brought up in the thread, as well as some things that may be particular to Japanese:
Of course it has SOV word order, adjective-noun, exclusively suffixing, nominative-accusative, agglutinating morphology.

It has case-marking (kind of): kao no, kotoba ni, ki o

There are two instances where relative clauses take special attributive forms:
conclusive attributive
kao ga mienai kao no mienai
can't see the face (whose) face can't be seen
Miyasalemmi does not have special attributive forms for relative clauses.

mishirazu mishiranu
is unknown (that) is unknown
ótúremána ótúremáne

verb compounding:
miru + shiru + nu mishiranu
see + know/realize + NEG.ATTR be.unknown
Not for this verb but for large amounts (I think even more pervasive than in J) others: (mil.senéra (repeat-think = recall); kuli.issa (love-unite = marry); tefti.déra (happiness-meet = party); mildéra (repeat-meet = reunite); haslemma (forceful-say = promise) ...)

locational comparative construction (that has been abbreviated to mean "more" and is actually the only instance I know of in Japanese where a particle does not follow anything)
ki o tsukeru yori ki o tsukeru
feeling/disposition OBJ attach from feeling/disposition OBJ attach
be careful be more careful
より only has an equivalent in Miyasalemmi when used as a clitic: それよりこれが好きです。

topic-comment structure (this topic is explicitly in the dative):
kotoba ni wa
word DAT TPC
lemmisi.hi.n.o = word.PLUR.DAT.TPC

Object promoted to subject in potential, evidence of the confusion of subject and object in Japanese:
kao o miru kao ga mieru
face OBJ see face SBJ see-POT
Miyasalemmi uses the passive potential in these cases, so face will get a subject marker.

idiomatic expression with ki:
ki o tsukeru
feeling/disposition OBJ attach
"be careful"
Miyasalemmi uses mimi=caution. mimikira = to hold caution = to take care

periphrastic expression for necessity (can be abbreviated to just the complementizer)
-nai to (ikenai)
NEG COMPL (go-can-NEG)
"must"
Miyasalemmi does not generally use periphrastic expressions. It goes like STEM-emánakyemán, roughly translated into Japanese as something like **見られなけれない

I tried my best. Idiomatic things don't translate easily but the big idea is the same. Miyasalemmi has this mechanism to make sure the scope of relative clauses is never ambiguous (it started as an attempt to fix/patch/fork Japanese, and eventually got relexed completely) so there are two particles added (nan for starting the relative clause, esti for linking the verbs into one clause).

[hwhatting]

Verbs in Altaic do not agree with their subjects for person and number (I don't know what the order of the morphemes that verbs do have is.).

Not true for Turkic - all Turkic languages I know have person and number marked on the verb, although Kazakh doesn't mark number on the 3rd person and, IIRC, in Uzbek number marking on the 3rd person is optional.

[gach]

Verbs in Altaic do not agree with their subjects for person and number (I don't know what the order of the morphemes that verbs do have is.).

Not true for Turkic - all Turkic languages I know have person and number marked on the verb, although Kazakh doesn't mark number on the 3rd person and, IIRC, in Uzbek number marking on the 3rd person is optional.

Also checking for what I could find for Tungusic, at least Udihe and Evenki including Solon definitely conjugate their finite verbs for person and number.

[kuroda]

Also checking for what I could find for Tungusic, at least Udihe and Evenki including Solon definitely conjugate their finite verbs for person and number.

Yes, definitely. All the Tungusic languages but Jurchen-Manchu have full or partial systems of person/number marking on verbs. Many Mongolian languages do or did, at different stages of their lives, as well.

In Tungusic languages, too, constituent order is generally much more free than in Turkic & Mongolian languages. (Again, except for Jurchen-Manchu.)

(Though, as a certain contemporary Manjurologist has told his students, "Manchu is Mongolian for dummies"... )

[sucaeyl]

Ok so I've done some work on this, and I'm becoming worried that I'm creating a re-lex of Japanese. How should I avoid this? What are some quirks of Japanese that are not representative of the other "Altaic" languages? Thanks!

Edit: I primarily mean morphosyntax, and to a lesser extent semantics. So yeah.

[kuroda]

To give a quick and dirty answer, if Japanese is the language you know best and are using as a default model of 'Altaic', then it might be easiest to get yourself a grammar of, say, Mongolian or Sakha or Kazakh and try to use _that_ as the basis for formulating morphosyntax et al. IOW, just start from a less familiar point and work towards what is

[Herr Dunkel]

Mongolian, Turkish, Japanese, by your powers combined I am Standard Average Altaic

[clawgrip]

Come on, Standard Average Altaic isn't that lame.