Linguistic Miscellany Thread

[Estav]

I've occasionally heard people use a preposition + which construction (does that have a name?) while also stranding the preposition at the end of the clause: "the store to which I went to", or something like that. But I recently heard someone say "the source from which it was influenced by", with two different prepositions!

Mark Liberman has written about these kinds of examples on Language Log: e.g http://itre.cis.upenn.edu/~myl/language ... 04465.html

[bradrn]

A couple of small questions which aren’t big enough for their own thread:

1. I was reading through a grammar of Chalcatongo Mixtec, and noticed that it uses completely different constructions for headless relative clauses and embedded questions:
   
   

   … xa-ní-žúbaʒa=rí

   COMP=CP-have=1

   
   … the one that I used to have / what I used to have
   
   

   … ndàsà

   how

   čaa=rí

   write=1

   
   … how to write
   
   By contrast, English merges these two forms: it has no construction used for headless relative clauses but not for embedded questions. How common is it for languages to have separate constructions for these forms as opposed to merging them?
2. Are there any languages with only one case-marker? (Or equivalently, are there any with two cases, one of which is realised as ∅?)
3. I’ve been reading Li and Thompson’s 1976 article about subject-prominent and topic-prominent languages. In it, they distinguish four types of language:
   

   According to our study, there are four basic types of languages: (i) languages that are subject-prominent … (ii) languages that are topic-prominent; (iii) languages that are both subject-prominent and topic-prominent; (iv) languages that are neither subject-prominent nor topic-prominent.

   In the article, they discuss types (i) and (ii) quite extensively. However, they don’t seem to comment on how languages of type (iii) or (iv) work. Does anyone have any more information on this?
4. As I was researching the previous question, I came across the following Standard Mandarin (I think) clause:
   
   丫头你提了个問題很嚴重，…
   

   Yātou

   girl

   nǐ

   2sg

   tí-le

   raise-PFV

   ge

   CL

   wèntí

   question

   hěn

   very

   yánzhòng, …

   serious

   
   Girl, you’ve raised a question that is very serious …
   
   Now, I’ve been reading about classifiers lately, and from what I understand of Standard Mandarin, classifiers are only ever used with numerals or deictics. However, here there is a classifier ge 个 being used before a noun with neither of these. Could someone more familiar with the language explain why that classifier is there?

[Kuchigakatai]

It's basically 一个 yíge with the yí removed; you can sometimes remove it but not always (the contexts where you can do that are determined by prosody). Yip and Rimmington's Comprehensive Grammar (2nd edition) has an eight-page discussion on prosody and 一个 yíge vs. 个 ge in pages 429-436.

[bradrn]

It's basically 一个 yíge with the yí removed; you can sometimes remove it but not always (the contexts where you can do that are determined by prosody). Yip and Rimmington's Comprehensive Grammar (2nd edition) has an eight-page discussion on prosody and 一个 yíge vs. 个 ge in pages 429-436.

That’s much more straightforward than I was thinking — thanks for the explanation!

But now I’m wondering why yí is there in the first place; when I look at the English translation, I can’t see any obvious reason why it should be there. That is, given two sentences:

丫头你提了一个問題很嚴重，…

Yātou

girl

nǐ

2sg

tí-le

raise-PFV

yí

one

ge

CL

wèntí

question

hěn

very

yánzhòng, …

serious

丫头你提了問題很嚴重，…

Yātou

girl

nǐ

2sg

tí-le

raise-PFV

wèntí

question

hěn

very

yánzhòng, …

serious

Why is the first preferred over the second?

[Kuchigakatai]

(一)个 (yí)ge clarifies that it's an indefinite singular. Otherwise in your second sentence the direct object could be understood as singular or plural, indefinite or definite.

In Mandarin it is actually very normal to mark an indefinite singular as such in the direct object position. It happens a lot more often than not, to the point that a bare dir. obj. noun phrase (as in your second sentence) generally receives an indefinite plural interpretation, and only secondarily the other three possible interpretations (def. pl., indef. sg., def. sg.).

If your direct object is definite, your sentence probably has the focus on an adverb(ial) / an indirect object / a resultative/potential/directional verb complement, which means you're probably using the 把 bǎ-construction anyway (not a regular direct object). It's not mandatory to use the 把 bǎ-construction with a definite dir. obj., but it is likely.

Differentiation of definiteness in direct objects that's otherwise absent in other syntactic positions is normal. Compare:
- the dir. obj. genitive in Russian, used if the verb is negated and the dir. obj. is indefinite, especially if inanimate; otherwise the accusative is used. Russian doesn't normally clarify definiteness elsewhere (except maybe by using pragmatic word order).
- the Finnish dir. obj. partitive case, used with indefinite mass nouns, or if the verb is negated (but it is also a true partitive meaning 'some of the X' if the noun is countable (i.e. not a mass noun))
- the dir. obj. definite marker râ in formal Persian, a language that otherwise does not mark definiteness in nouns with a postposition (amusingly, râ is also used with personal pronouns like 'me' and 'you' --isn't a pers. pron. obviously definite?)
- the "personal a" in Spanish, which is only mandatory if the human dir. obj. is definite (also used with personal pronouns, but at least Spanish has a free-er constituent order than Persian...). Spanish otherwise doesn't distinguish definiteness by adding/removing prepositions.

[dewrad]

Are there any languages with only one case-marker? (Or equivalently, are there any with two cases, one of which is realised as ∅?)

This post is written in such a language.

[Zju]

Are there languages with different (sets of) demonstrative pronouns depending on if they're a NP head? E.g. a language in which 'this' in 'I see this' and 'I see this book' would be translated differently.

[Ares Land]

Are there languages with different (sets of) demonstrative pronouns depending on if they're a NP head? E.g. a language in which 'this' in 'I see this' and 'I see this book' would be translated differently.

French. Je vois ça vs Je vois ce livre(-là) or Je vois cela vs. je vois ce livre(-là) though the latter is kind of dated.

[Kuchigakatai]

Are there languages with different (sets of) demonstrative pronouns depending on if they're a NP head? E.g. a language in which 'this' in 'I see this' and 'I see this book' would be translated differently.

I think that first of all it'd be important to distinguish the determiner 'this book' (pl. these books), the (rather abstract) pronoun 'this', and the nominalized determiner 'this one' (pl. these ones, also expressable as 'these').

Spanish technically counts, but only because the pronoun 'this' is different (esto); otherwise its equivalent of 'this one' is a pronoun that sounds the same as the determiner of 'this [book]' (este/esta/estos/estas). Esto veo 'I see this', veo este libro 'I see this book', este veo 'I see this one' (this masculine thing: el libro).

French is a much much better example:
- pronoun 'this' -> ça /sɑ/, formally also ceci /səsi/
- 'this [book]' -> ce/cet/cette/ces pronounced /sə sət sɛt se/ (formally ces = /sɛ/), also colloquial French ç'te/ç't' pronounced /stə st/ for ce/cet
- 'this one' -> celui-là /səlɥilɑ/ (and {celle,ceux,celles}-là pronounced /sɛl,sø,sɛl + lɑ/), colloquial French also çui-là /sɥilɑ/ for celui-là, formally also celui-ci /səlɥisi/ (and {celle,ceux,celles}-ci), besides phrases like

Je vois ceci. ~ Je vois ça. /ʒəvwasəsi ~ ʒvwasɑ/
'I see this.'

Je vois ce livre-ci. ~ Je vois ce livre. ~ Je vois ç'te livre. /ʒəvwɑ səlivʁəsi ~ ʒvwɑ s(ə)liv(ʁ) ~ ʒvwɑ stəliv(ʁ)/
'I see this book.'

Je vois celui-ci. ~ Je vois celui-là. ~ Je vois çui-là. /ʒəvwɑ səlɥisi ~ ʒvwɑ s(ə)lɥilɑ ~ ʒvwɑ sɥilɑ/
'I see this one.'


(Putting it in before our other French speakers complain: yes, I know mentioning çui-là is possibly pushing it a bit too much, but ehhhh, it seems arguably common enough to be worth mentioning? I understand any complaints though.)

[Ares Land]

(Putting it in before our other French speakers complain: yes, I know mentioning çui-là is possibly pushing it a bit too much, but ehhhh, it seems arguably common enough to be worth mentioning? I understand any complaints though.)

No, it's completely legit. Spoken French is just as valid as written French for our purpose, and [səlɥilɑ] is only used for emphasis or in formal contexts.

[Richard W]

Are there any languages with only one case-marker? (Or equivalently, are there any with two cases, one of which is realised as ∅?)

This post is written in such a language.

On reading the question, I assumed this idea was rejected because pronouns behave differently, e.g. with the accusative suffix -m. Additionally, there's a feeling that English-s is a postposition.

The Persian definite object suffix را /rɑ/ is the only example I can think of for sure, as my knowledge of pronouns is more limited.

[bradrn]

(一)个 (yí)ge clarifies that it's an indefinite singular. Otherwise in your second sentence the direct object could be understood as singular or plural, indefinite or definite.

In Mandarin it is actually very normal to mark an indefinite singular as such in the direct object position. It happens a lot more often than not, to the point that a bare dir. obj. noun phrase (as in your second sentence) generally receives an indefinite plural interpretation, and only secondarily the other three possible interpretations (def. pl., indef. sg., def. sg.).

If your direct object is definite, your sentence probably has the focus on an adverb(ial) / an indirect object / a resultative/potential/directional verb complement, which means you're probably using the 把 bǎ-construction anyway (not a regular direct object). It's not mandatory to use the 把 bǎ-construction with a definite dir. obj., but it is likely.

Differentiation of definiteness in direct objects that's otherwise absent in other syntactic positions is normal. Compare:
- the dir. obj. genitive in Russian, used if the verb is negated and the dir. obj. is indefinite, especially if inanimate; otherwise the accusative is used. Russian doesn't normally clarify definiteness elsewhere (except maybe by using pragmatic word order).
- the Finnish dir. obj. partitive case, used with indefinite mass nouns, or if the verb is negated (but it is also a true partitive meaning 'some of the X' if the noun is countable (i.e. not a mass noun))
- the dir. obj. definite marker râ in formal Persian, a language that otherwise does not mark definiteness in nouns with a postposition (amusingly, râ is also used with personal pronouns like 'me' and 'you' --isn't a pers. pron. obviously definite?)
- the "personal a" in Spanish, which is only mandatory if the human dir. obj. is definite (also used with personal pronouns, but at least Spanish has a free-er constituent order than Persian...). Spanish otherwise doesn't distinguish definiteness by adding/removing prepositions.

Thanks!

Are there any languages with only one case-marker? (Or equivalently, are there any with two cases, one of which is realised as ∅?)

This post is written in such a language.

Unless I’m badly misunderstanding something, English doesn’t have any casemarkers — the cases are fused with pronouns.

Are there languages with different (sets of) demonstrative pronouns depending on if they're a NP head? E.g. a language in which 'this' in 'I see this' and 'I see this book' would be translated differently.

Dixon’s Basic Linguistic Theory has a lovely table summarising this. Answering your question:

• Some languages (e.g. French, Japanese) have different forms for all of this, this (modifier), here
• Some languages (e.g. Mangap-Mbula) have the same form for this (modifier) and here, but a different one for this
• Some languages (e.g. Ponapean, Lango) have this (modifier) as a bound affix

(Of course this is not an exhaustive list — I only listed the languages which were applicable to your question.)

[dewrad]

Are there any languages with only one case-marker? (Or equivalently, are there any with two cases, one of which is realised as ∅?)

This post is written in such a language.

Unless I’m badly misunderstanding something, English doesn’t have any casemarkers — the cases are fused with pronouns.

Leaving pronouns aside, English NPs mark two cases: possessive and non-possessive. That the possessive is marked by a phrase-level clitic is really neither here nor there (if you exclude clitics as case markers, then Sumerian must be considered a caseless language, for example).

If English doesn’t suit, however, there’s always Old French and Old Occitan, which both have two cases (direct and oblique), which share a single case-marker: -s marks the singular direct as well as the plural oblique.

[bradrn]

This post is written in such a language.

Unless I’m badly misunderstanding something, English doesn’t have any casemarkers — the cases are fused with pronouns.

Leaving pronouns aside, English NPs mark two cases: possessive and non-possessive. That the possessive is marked by a phrase-level clitic is really neither here nor there (if you exclude clitics as case markers, then Sumerian must be considered a caseless language, for example).

If English doesn’t suit, however, there’s always Old French and Old Occitan, which both have two cases (direct and oblique), which share a single case-marker: -s marks the singular direct as well as the plural oblique.

Thanks! That Old French system in particular looks like what I was looking for. (I was thinking of giving my conlang a case-marking system with only one case-marker for the ergative.)

(And by the way, I do definitely consider adpositions and clitics to be case-markers: in fact, the aforementioned ergative case-marker is going to be a clitic.)

[bradrn]

It looks like my other two questions have gotten a bit lost in all the discussion, so I’ll re-post them:

1. I was reading through a grammar of Chalcatongo Mixtec, and noticed that it uses completely different constructions for headless relative clauses and embedded questions:
   
   

   … xa-ní-žúbaʒa=rí

   COMP=CP-have=1

   
   … the one that I used to have / what I used to have
   
   

   … ndàsà

   how

   čaa=rí

   write=1

   
   … how to write
   
   By contrast, English merges these two forms: it has no construction used for headless relative clauses but not for embedded questions. How common is it for languages to have separate constructions for these forms as opposed to merging them?
2. I’ve been reading Li and Thompson’s 1976 article about subject-prominent and topic-prominent languages. In it, they distinguish four types of language:
   

   According to our study, there are four basic types of languages: (i) languages that are subject-prominent … (ii) languages that are topic-prominent; (iii) languages that are both subject-prominent and topic-prominent; (iv) languages that are neither subject-prominent nor topic-prominent.

   In the article, they discuss types (i) and (ii) quite extensively. However, they don’t seem to comment on how languages of type (iii) or (iv) work. Does anyone have any more information on this?

[Richard W]

By contrast, English merges these two forms: it has no construction used for headless relative clauses but not for embedded questions. How common is it for languages to have separate constructions for these forms as opposed to merging them?

I'm pretty sure that the commonality in English arises because English has largely merged relative and interrogative pronouns. I thin that is a peculiarly Western IE thing, so inherently rare.

[Kuchigakatai]

Thanks!

Oh, and if you have a definite direct object in Mandarin, you might also topicalize it by moving it to the beginning, or even omit it as understood in the context.

This post is written in such a language.

Unless I’m badly misunderstanding something, English doesn’t have any casemarkers — the cases are fused with pronouns.

Leaving pronouns aside, English NPs mark two cases: possessive and non-possessive. That the possessive is marked by a phrase-level clitic is really neither here nor there (if you exclude clitics as case markers, then Sumerian must be considered a caseless language, for example).

And even then, the Old English genitive has not been completely kicked out of words as a clitic. People still say things like "I liked Michael's and Jamie's work" while referring to a single piece of work the two guys accomplished together, pretty much treating -'s like a typical genitive of Latin or Russian (Milōnis Jūniīque opus). Not everyone finds "the man that I saw yesterday's hat" perfectly felicitous, and I'd say that even familiarity with the academic "Jones et al.'s work on Gaulish" probably requires a peculiar exposure (rather than being wholly natural in English)... I wonder to what extent people actually say "Consult Baxter and others' work". Surely some do, but...

By the way, if anyone here knows, what are other known instances of inflectional morphemes pretty much undergoing de-grammaticalization, as it happened (or it is currently happening) to the Indo-European genitive ending in English? This kind of thing seems to be really uncommon.

In Spanish you can attach the plural morpheme to phrases once they're phonologically and semantically tight enough. Corre, ve y dile 'run, see and tell him/her' > el correveidile 'tattler, snitch', plural los correveidiles. The RAE says the word for 'sofa-bed' is written <sofá cama> representing /soˈfa ˈkama/, pluralized sofás cama or sofás camas, but the reality is that many people also pronounce it /sofaˈkama/ (el sofacama) and pluralize it as /sofaˈkamas/ (los sofacamas). But this has more to do with phrases becoming compounds (as unusual as that is in Spanish) than any presumed separation (and increasing scope) of the plural morpheme.

Something a bit related, except not really, is the thing Polish did, attaching the regular personal verbal agreement endings at the end of participles, which is... unusual. Now just imagine Spanish forming its past tense that way: yo amado/amádao (masc./fem.), tú amádoas/amadas, nosotros amádosmos, ellos amádosan...

[Pabappa]

i like the Polish setup .... Ive been wanting to use that in a conlang for a long time but the languages I have that use gender are so un-IE-like that it doesnt really feel the same. I know when I was young I thought Spanish marked 3rd person verbs with -o -a -os -as instead of -a/-an ... it didnt occur to me that that would conflict with the 1st person. but my mistake was seeing a participle and mistaking it for a verb. something like that could happen if two forms happen to collide. though i think in Polish there might have been a padding morpheme between the participle and the verb ending, with the meaning "be" ... but i dont know. Russian has a halfway system where it uses participles that mark gender as its past tense verbs but does not put person endings on.

on your other question .... english 'll could be considered an inflectional mropheme, but because its just a contraction of will, its not really the same .... "the dog that's here'll go to sleep when you do" is not as impressive as "the dog that's here's blanket is in the wash" etc.

edit: oh by the way, i love correveidile. there dont seem to be too many Spanish words like that.

[Travis B.]

Thanks!

Oh, and if you have a definite direct object in Mandarin, you might also topicalize it by moving it to the beginning, or even omit it as understood in the context.

Unless I’m badly misunderstanding something, English doesn’t have any casemarkers — the cases are fused with pronouns.

Leaving pronouns aside, English NPs mark two cases: possessive and non-possessive. That the possessive is marked by a phrase-level clitic is really neither here nor there (if you exclude clitics as case markers, then Sumerian must be considered a caseless language, for example).

And even then, the Old English genitive has not been completely kicked out of words as a clitic. People still say things like "I liked Michael's and Jamie's work" while referring to a single piece of work the two guys accomplished together, pretty much treating -'s like a typical genitive of Latin or Russian (Milōnis Jūniīque opus). Not everyone finds "the man that I saw yesterday's hat" perfectly felicitous, and I'd say that even familiarity with the academic "Jones et al.'s work on Gaulish" probably requires a peculiar exposure (rather than being wholly natural in English)... I wonder to what extent people actually say "Consult Baxter and others' work". Surely some do, but...

By the way, if anyone here knows, what are other known instances of inflectional morphemes effectively going de-grammaticalization, as it happened (or it is currently happening) to the Indo-European genitive ending in English? This kind of thing seems to be really uncommon.

To my knowledge, the very same thing that happened to the genitive in English also happened to the genitive in continental Scandinavian.

[bradrn]

By contrast, English merges these two forms: it has no construction used for headless relative clauses but not for embedded questions. How common is it for languages to have separate constructions for these forms as opposed to merging them?

I'm pretty sure that the commonality in English arises because English has largely merged relative and interrogative pronouns. I thin that is a peculiarly Western IE thing, so inherently rare.

Ah, right. So most non-IE languages keep these constructions separate.

But actually, now that I think about it: how many non-IE languages have a embedded questions in the first place?

Not everyone finds "the man that I saw yesterday's hat" perfectly felicitous, and I'd say that even familiarity with the academic "Jones et al.'s work on Gaulish" probably requires a peculiar exposure (rather than being wholly natural in English)... I wonder to what extent people actually say "Consult Baxter and others' work". Surely some do, but...

I find "Jones et al.'s work on Gaulish" and "Consult Baxter and others' work" far more natural than the alternatives. But "the man that I saw yesterday's hat" is certainly not quite felicitous for me; I would be more likely to say “the hat of the man that I saw yesterday”.

Something a bit related, except not really, is the thing Polish did, attaching the regular personal verbal agreement endings at the end of participles, which is... unusual. Now just imagine Spanish forming its past tense that way: yo amado/amádao (masc./fem.), tú amádoas/amadas, nosotros amádosmos, ellos amádosan...

This reminds me of Walmatjari, which I was reading about yesterday. Walmatjari marks personal agreement and mood on an obligatory auxiliary rather than the verb itself:

yani marna I went
yani pajarra We 2 went
yani pa He went
yani pila They two went
etc.

(Note that the verb here is yani; the auxilliary is stated as usually being the second word in the sentence.)