Zompist Bboard Again

Alright, clearly my original question was horribly vague and didn’t get me the sort of answers I was looking for. So let me re-ask my question in a different form: Is there any language in which articles appear closer to the noun than adjectives or relative clauses? Hopefully that should be unambiguous enough that no-one can poke any holes in it…

zompist wrote: ↑Sat Aug 01, 2020 10:22 pm In syntax, we say that (English) articles and quantifiers both live in the Det node, which always precedes the N' that makes up the rest of the NP.

This is why I decided to ask that question in the first place: my (limited) understanding of syntax was that the determiner is always outside the NP, giving a DetP. Hence my asking if it’s possible for the determiner to be inside the NP.

Biblical Hebrew has some interesting rules regarding the construct state, which is used for expressions like "the word of the king", "the god of Israel", "the house of peace", etc. The first word is in a distinct morphological form, the contruct state. The interesting bit is that the article can apply only to the other word, and makes both words definite. Thus dəbar ha-mmelek "the word of the king" where the article ha- attaches to the second (non-construct) word, though it is semantically only a modifier to the first.

Interesting… but how is ha- semantically a modifier to the first word? It’s very common for possessed nouns to be automatically definite; to me this seems like such a case, where ha- affects the definiteness of the possessor melek, while the possessed word dəbar is always definite.

Isn't the noun in the construct state definite regardless?

bradrn wrote: ↑Sat Aug 01, 2020 10:41 pm Alright, clearly my original question was horribly vague and didn’t get me the sort of answers I was looking for. So let me re-ask my question in a different form: Is there any language in which articles appear closer to the noun than adjectives or relative clauses? Hopefully that should be unambiguous enough that no-one can poke any holes in it…

No. In 'the house that Jack built', the article feels closer even though you might draw a tree structure showing the opposite. What about

The house fell down that we lived in ten years ago.

? That sort of structure happens - big relative clauses get shoved to the end because they just confuse matters in the middle of a sentence.

Richard W wrote: ↑Sat Aug 01, 2020 10:53 pm

bradrn wrote: ↑Sat Aug 01, 2020 10:41 pm Alright, clearly my original question was horribly vague and didn’t get me the sort of answers I was looking for. So let me re-ask my question in a different form: Is there any language in which articles appear closer to the noun than adjectives or relative clauses? Hopefully that should be unambiguous enough that no-one can poke any holes in it…

No.

Really? There are no languages at all in which this happens? I can’t say that that’s unexpected, but I think I’d prefer a source for a blanket statement like that.

In 'the house that Jack built', the article feels closer even though you might draw a tree structure showing the opposite.

Yes, but ‘feels closer’ doesn’t really mean anything. I suppose I’m looking more for languages which have structures like ‘big blue the house’ or ‘house the that Jack build’, where the article is unambiguously closer than the modifiers to the noun.

What about

The house fell down that we lived in ten years ago.

?

That’s ungrammatical for me.

Richard W wrote: ↑Sat Aug 01, 2020 10:44 pm Isn't the noun in the construct state definite regardless?

No, you can have construct expressions like"a word of a king"-- I believe that's dəbar melek. (Hebrew letters are on the Mac...)

You may be thinking of the rule that proper nouns are always definite. E.g bet Dawid "the house of David" is definite though there's no article on "house".

bradrn wrote: ↑Sat Aug 01, 2020 10:41 pm Alright, clearly my original question was horribly vague and didn’t get me the sort of answers I was looking for. So let me re-ask my question in a different form: Is there any language in which articles appear closer to the noun than adjectives or relative clauses? Hopefully that should be unambiguous enough that no-one can poke any holes in it…

The Turkish indefinite article bir occurs between adjectives and the noun. (If bir is put before an adjective, it has to be interpreted as the number one.) I wouldn't be surprised if the same thing with definite articles is rare or even nonexistent, though.

akam chinjir wrote: ↑Sun Aug 02, 2020 12:52 am The Turkish indefinite article bir occurs between adjectives and the noun. (If bir is put before an adjective, it has to be interpreted as the number one.)

Thanks!

I wouldn't be surprised if the same thing with definite articles is rare or even nonexistent, though.

How come? I know there are plenty of differences between the semantics of definite and indefinite articles, but I had imagined that their syntax would at least be comparable.

bradrn wrote: ↑Sun Aug 02, 2020 1:09 am How come? I know there are plenty of differences between the semantics of definite and indefinite articles, but I had imagined that their syntax would at least be comparable.

I think indefinite articles are generally thought to be a part of the NP, whereas a definite article takes an NP as its complement. (But maybe I'm wrong about this? I've read vanishingly little about indefinite articles, and nothing syntax-y about the Turkish ones.)

akam chinjir wrote: ↑Sun Aug 02, 2020 1:14 am

bradrn wrote: ↑Sun Aug 02, 2020 1:09 am How come? I know there are plenty of differences between the semantics of definite and indefinite articles, but I had imagined that their syntax would at least be comparable.

I think indefinite articles are generally thought to be a part of the NP, whereas a definite article takes an NP as its complement. (But maybe I'm wrong about this? I've read vanishingly little about indefinite articles, and nothing syntax-y about the Turkish ones.)

That’s an interesting analysis — I would have assumed that both articles are analysed the same way. Why are they analysed differently?

bradrn wrote: ↑Sun Aug 02, 2020 1:17 am That’s an interesting analysis — I would have assumed that both articles are analysed the same way. Why are they analysed differently?

Not sure, and now I'm full of doubts

zompist wrote: ↑Sat Aug 01, 2020 10:22 pmBiblical Hebrew has some interesting rules regarding the construct state, which is used for expressions like "the word of the king", "the god of Israel", "the house of peace", etc. The first word is in a distinct morphological form, the contruct state. The interesting bit is that the article can apply only to the other word, and makes both words definite. Thus dəbar ha-mmelek "the word of the king" where the article ha- attaches to the second (non-construct) word, though it is semantically only a modifier to the first.

zompist wrote: ↑Sat Aug 01, 2020 11:22 pmNo, you can have construct expressions like"a word of a king"-- I believe that's dəbar melek. (Hebrew letters are on the Mac...)

I don't remember what Hebrew does exactly, and it's soon going to be midnight here and I don't have time to check, but judging by what Classical Arabic does, wouldn't it be more accurate to say that the article attached to the second noun changes the definiteness of both nouns at the same time? So your only possibilities would only be "a word of a king" and "the word of the king", and other permutations would use other constructions. Arabic is not Hebrew, but you seem to be kind of suggesting this.

(In Classical Arabic, "the word of a king" and "a word of the king" would use a prepositional phrase for the possessor, not the construct state + genitive construction.)

bradrn wrote: ↑Sat Aug 01, 2020 11:00 pm

Richard W wrote: ↑Sat Aug 01, 2020 10:53 pmWhat about

The house fell down that we lived in ten years ago.

?

That’s ungrammatical for me.

That's a common construction in old Indo-European languages (Old English, Latin, Ancient Greek, Sanskrit...). I think that's what Richard W was alluding to. You can still find this sort of construction, kind of, in proverbs in modern English, e.g. "He is wise who can make a friend of a foe" *, where the relative clause is still modifying "he", arguably.

* Personally, I like the version often ascribed to Abraham Lincoln, as I feel it delivers the point better: "Do I not destroy my enemies when I make them my friends?".

Ser wrote: ↑Sun Aug 02, 2020 1:52 am

zompist wrote: ↑Sat Aug 01, 2020 10:22 pmBiblical Hebrew has some interesting rules regarding the construct state, which is used for expressions like "the word of the king", "the god of Israel", "the house of peace", etc. The first word is in a distinct morphological form, the contruct state. The interesting bit is that the article can apply only to the other word, and makes both words definite. Thus dəbar ha-mmelek "the word of the king" where the article ha- attaches to the second (non-construct) word, though it is semantically only a modifier to the first.

zompist wrote: ↑Sat Aug 01, 2020 11:22 pmNo, you can have construct expressions like"a word of a king"-- I believe that's dəbar melek. (Hebrew letters are on the Mac...)

I don't remember what Hebrew does exactly, and it's soon going to be midnight here and I don't have time to check, but judging by what Classical Arabic does, wouldn't it be more accurate to say that the article attached to the second noun changes the definiteness of both nouns at the same time? So your only possibilities would only be "a word of a king" and "the word of the king", and other permutations would use other constructions. Arabic is not Hebrew, but you seem to be kind of suggesting this.

So far as I can see, that's exactly what I said. Yes, you get either "the word of the king" or "a word of a king".

I think it's relevant to Brad's original question simply because definiteness is here not marked on the head at all. (Assuming "word" is the head... I'm pretty sure it governs agreement, so it probably is...)

bradrn wrote: ↑Sat Aug 01, 2020 11:00 pm

Richard W wrote: ↑Sat Aug 01, 2020 10:53 pm

bradrn wrote: ↑Sat Aug 01, 2020 10:41 pm Alright, clearly my original question was horribly vague and didn’t get me the sort of answers I was looking for. So let me re-ask my question in a different form: Is there any language in which articles appear closer to the noun than adjectives or relative clauses? Hopefully that should be unambiguous enough that no-one can poke any holes in it…

No.

Really? There are no languages at all in which this happens? I can’t say that that’s unexpected, but I think I’d prefer a source for a blanket statement like that.

That's no to the unambiguity.

Balto-Slavonic may provide a counter-example. The long-form adjectives embed a demonstrative (and some aspects of the separate inflection of adjective and article survives to this day), which appears to imply adjective-article-noun. However, it would be good to have confirmation that OCS conforms to this pattern; I've encountered some odd statements when trying to verify this.

Ser wrote: ↑Sun Aug 02, 2020 1:52 am [hat's a common construction in old Indo-European languages (Old English, Latin, Ancient Greek, Sanskrit...). I think that's what Richard W was alluding to. You can still find this sort of construction, kind of, in proverbs in modern English, e.g. "He is wise who can make a friend of a foe" *, where the relative clause is still modifying "he", arguably.

I occasionally utter the displaced relative clause myself. It is a bit awkward, but following the normal word order results in an even more awkward sentence. There are other tricks, but they make for an explicitly disjointed sentence.

Postscript:
Googling for the word sequence "burnt down that we", I found a few examples:

Lack of activity and sedentary lifestyle is major result which results in calories not being burnt down that we have consumed.

I can't blame her, since the house had just burnt down that we were going to live in together, you know.

akam chinjir wrote: ↑Sun Aug 02, 2020 1:21 am

bradrn wrote: ↑Sun Aug 02, 2020 1:17 am That’s an interesting analysis — I would have assumed that both articles are analysed the same way. Why are they analysed differently?

Not sure, and now I'm full of doubts

One possible explanation for such a difference is that the definite article commonly derives from demonstratives and indefinite articles often from the cardinal 'one', and demonstratives and quantifiers may occupy different slots, e.g. noun-classifier-demonstrative and noun-numeral-classifier in Thai. On the other hand, the Thai word for 'one' apparently used as an indefinite article is reported to occupy the demonstrative slot, as does the 'one only' word (เดียว /diəu/). Matters seem chaotic with respect to the 'one' word.

Richard W wrote: ↑Sun Aug 02, 2020 5:16 amI occasionally utter the displaced relative clause myself. It is a bit awkward, but following the normal word order results in an even more awkward sentence. There are other tricks, but they make for an explicitly disjointed sentence.

Postscript:
Googling for the word sequence "burnt down that we", I found a few examples:

Lack of activity and sedentary lifestyle is major result which results in calories not being burnt down that we have consumed.

I can't blame her, since the house had just burnt down that we were going to live in together, you know.

They sound pretty ungrammatical to me... The first example there looks very much like L2 writing too, with its use of "major result" for what I imagine was intended to be "common habit" unless its context really makes it appropriate, and the lack of an indef. art. before that ("is a [major result]"). Maybe you've been learning Pali for too long.

zompist wrote: ↑Sun Aug 02, 2020 2:13 amSo far as I can see, that's exactly what I said. Yes, you get either "the word of the king" or "a word of a king".

I think it's relevant to Brad's original question simply because definiteness is here not marked on the head at all. (Assuming "word" is the head... I'm pretty sure it governs agreement, so it probably is...)

I found that Biblical Hebrew does use a prepositional phrase (with לְ lə-) if the possessed noun is indefinite and the possessor is definite (Muraoka's Joüon gives the example נָבִיא ליהוה nāḇîʾ l-YHWH 'a prophet of God'), but I continue to struggle to find about the opposite case.

Regarding agreement, both the possessed noun and the possessor noun can govern adjectives, which appear definite or indefinite as appropriate and agreeing in gender+number:
דִּבְרֵי דָּוִד הָאַחֲרֹנִים diḇrê dāwiḏ hā-ʾaḥărōnîm 'the last words of David' (2 Samuel 23:1)
כִּדְגַת הַיָּם הַגָּדוֹל kiḏəḡaṯ hay-yām hag-gāḏôl 'like the fish of the Great Sea (the Mediterranean)' (Ezekiel 47:10)

Unrelatedly, while trying to find more information about this, I learned that Biblical Hebrew has a construction similar to the Arabic "false annexation/ʔidˤaafa", where an adjective takes the construct state and is modified by a genitive possessor to form an overall adjectival phrase:
אֶ֫רֶץ זָבַת חָלָב וּדְבַשׁ ʾe´reṣ zāḇaṯ ḥālāḇ û-ḏəḇaš 'a land flowing with milk and honey', in which the participial זָבָה zāḇāh 'flowing' appears in the construct state (this also happens with normal adjectives, not just participles).

I have a conlang where genitives are formed like possessee GEN possessor. But in endocentric compounds (including noun-noun compounds) the head is the last stem. Is this wrong? It seems like the genitive construction and compound words go in opposite directions.

not really. english sort of does that, right? "book case" vs "case of books". and likewise a lot of other IE languages and probably non-IE ones. it might even be easier to communicate this way since a listener is more likely to hear the difference between two phrases if the missing piece is not just a small, unstressed case marker.

Richard W wrote: ↑Sun Aug 02, 2020 4:54 am

bradrn wrote: ↑Sat Aug 01, 2020 11:00 pm

Richard W wrote: ↑Sat Aug 01, 2020 10:53 pm
No.

Really? There are no languages at all in which this happens? I can’t say that that’s unexpected, but I think I’d prefer a source for a blanket statement like that.

That's no to the unambiguity.

Balto-Slavonic may provide a counter-example. The long-form adjectives embed a demonstrative (and some aspects of the separate inflection of adjective and article survives to this day), which appears to imply adjective-article-noun. However, it would be good to have confirmation that OCS conforms to this pattern; I've encountered some odd statements when trying to verify this.

I dunno about long adjectives, but articles could appear either at the begining of the NP, or in its Wackernagel position (that is, N art, ADJ art N, ADJ art ADJ N, etc). Arguably the mid-distal demonstrative had already begun acquiring semantics of a definite article.

More samples:

... in 1983 our house got burnt down that we were living in, in Geraldton ... (unless that's Australian Aboriginal)

Completed September 2012, The Woolshed replaces an the old homestead that had burnt down that they had operated in. - Kiwi

For those guys to live for all these years, to see their building burnt down that they put time and effort into, that was huge.

... but unfortunately their Minivan burnt down which they just bought ...

And here's a paper on the usage - When relative clause extraposition is the right choice, it’s easier. I haven't read it - I was only after an existence proof.