Zompist Bboard Again

I just noticed this on Twitter:

https://twitter.com/alon_levy/status/10 ... 5465881601

My translation of the Hebrew tweet is almost perfectly word for word, but I just noticed something important about the original Hebrew syntax: the original tweet says הרבנות יפסיקו, ha-Rabbanut yafsiku, the-Rabbinate stop-fut.-3pl. In English, treating various large corporate bodies as plural, like the police, the government, or a large corporation is common: Facebook assured me that their intentions are good, the police say that they are effective at fighting crime, etc. In Hebrew, it is not. Abstracted nouns referring to collectives of people are always singular and take singular agreement; Hebrew speakers who learn English are reminded to use plural agreement for the police, because it never happens in Hebrew. And yet, the original tweep borrowed English syntax.

Are there good examples elsewhere for this process? Perhaps parallel evolution of aspect in Western European languages, using "have" to denote the perfect?

The European have perfect (and indeed possibly the be perfect) certainly looks like a prime candidate for borrowing, though as far as I know nobody has any firm evidence-based proposals for what the pathways of borrowing might have been, and it is possible (if not very appealing) to explain it as a number of separate language-internal processes.

Alon, you might be interested in the concept of Sprachbünde or linguistic areas: https://en.wikipedia.org/wiki/Sprachbund

Alon wrote: ↑Wed Jan 16, 2019 6:24 amIn English, treating various large corporate bodies as plural, like the police, the government, or a large corporation is common: Facebook assured me that their intentions are good, the police say that they are effective at fighting crime, etc.

You're not wrong. Treating a corporate body acting in disunity as plural is proper English: "The council were not in agreement." However, all of the examples you gave sound like "gender-nonspecific they" to me.

To answer your question, though, yes, syntactic borrowing can absolutely happen. Some prominent examples include Akkadian becoming SOV (originally VSO) under the influence of Sumerian, some modern Maya languages becoming SVO (originally OSV) under the influence of Spanish, Armenian becoming agglutinating under the influence of its Caucasian neighbors, some Indo-Iranian languages developing an ergative alignment, etc.

Zaarin wrote: ↑Wed Jan 16, 2019 10:47 amTo answer your question, though, yes, syntactic borrowing can absolutely happen. Some prominent examples include Akkadian becoming SOV (originally VSO) under the influence of Sumerian, some modern Maya languages becoming SVO (originally OSV) under the influence of Spanish, Armenian becoming agglutinating under the influence of its Caucasian neighbors, some Indo-Iranian languages developing an ergative alignment, etc.

Well that first case is predicated on the assumption that Proto-Semitic was VSO to begin with, which in my opinion doesn't actually hold much water. Secondly Maya was never OSV, is that a typo for VSO/VOS? Furthermore I don't see any areal element to the development of ergativity in Indo-Iranian: aside from maybe Burushaski I guess? there's no language group which could have done this - Sumerian was long dead before any Iranian language acquired the pattern, and neither Munda nor Dravidian are ergative so Indic can't have got their from there.

The point as a whole still stands though.

Alon wrote: ↑Wed Jan 16, 2019 6:24 amAre there good examples elsewhere for this process? Perhaps parallel evolution of aspect in Western European languages, using "have" to denote the perfect?

On the subject of perfects, Irish has a so-called "perfect" construction despite lacking any verb of possession:

Tá an t-alt san scríofa agam. "I've written that article."

Note the similarity to the common periphrastic expression of possession:

Tá an t-alt san agam. "I have that article." (Lit. "That article is at-me".)

In traditional Irish, this construction could only be used for the recent past. But I've noticed it being used more and more to translate the English perfect even when it refers to the distant past (e.g. Is beag alt atá scríofa agam "Few are the articles I've written"). This seems to be a pretty clear case of L1 interference.

Frislander wrote: ↑Wed Jan 16, 2019 11:06 am

Zaarin wrote: ↑Wed Jan 16, 2019 10:47 amTo answer your question, though, yes, syntactic borrowing can absolutely happen. Some prominent examples include Akkadian becoming SOV (originally VSO) under the influence of Sumerian, some modern Maya languages becoming SVO (originally OSV) under the influence of Spanish, Armenian becoming agglutinating under the influence of its Caucasian neighbors, some Indo-Iranian languages developing an ergative alignment, etc.

Well that first case is predicated on the assumption that Proto-Semitic was VSO to begin with, which in my opinion doesn't actually hold much water. Secondly Maya was never OSV, is that a typo for VSO/VOS? Furthermore I don't see any areal element to the development of ergativity in Indo-Iranian: aside from maybe Burushaski I guess? there's no language group which could have done this - Sumerian was long dead before any Iranian language acquired the pattern, and neither Munda nor Dravidian are ergative so Indic can't have got their from there.

The point as a whole still stands though.

Regarding Semitic, yes, it's based on the assumption that PS was VSO, which seems to be supported by the fact that not only are most Semitic languages VSO but so was their cousin Egyptian. With Akkadian being the odd language out and with its syntax resembling Sumerian (and other languages of Mesopotamia), I think it's fair to attribute the change to areal influence.

Regarding Maya, yes that was a typo for VOS.

I mentioned Indo-Iranian because ergativity is not typologically normal for IE languages and because a number of ergative languages are in the area that could have influenced it (including Hurro-Urartian and the languages of the Caucasus), though the only II languages I'm even passingly familiar with are Avestan and Farsi (and I know little of their grammar or syntax) so this was indeed an assumption on my part.

But split ergativity only evolved recently in Indo-Aryan, in the early to mid second millennium, not even reaching unrecognized languages that are considered dialects of Hindi, like Bhojpuri.

Are South Semitic languages mainly VSO either historically or currently?
Also, I'm not sure about the relationship between Semitic and Egyptian

mèþru wrote: ↑Thu Jan 17, 2019 6:54 am Are South Semitic languages mainly VSO either historically or currently?
Also, I'm not sure about the relationship between Semitic and Egyptian

Basically the only groups that are or have been VSO are Northwest Semitic and Arabic and its close relatives. Modern South Arabian is pretty universally SVO afaict, and Ethiosemitic universally SOV except for Ge‘ez, and there's reason to think they might have been intentionally mirroring the verb-initial syntax of its northern neighbours (in fact imho Ge‘ez suffers generally from people projecting onto it from a Proto-Semitic that had already been reconstructed based on Asian Semitic, which is just the wrong war round tbh). Additionally Ethiosemitic is a much deeper and more diverse branch of the family than many people realise, to the extent that it might actually be better to posit the PS Urheimat in the horn of Africa. Furthermore I haven't seen anything that proves that Semitic as a whole is specially close to Egyptian within Afro-Asiatic, and more distant from Cushitic.

I don't think Afroasiatic is valid actually. Perhaps a relationship between Semitic and Berber, but I doubt the other supposed branches are related to Semitic.

mèþru wrote: ↑Thu Jan 17, 2019 6:54 amAre South Semitic languages mainly VSO either historically or currently?
Also, I'm not sure about the relationship between Semitic and Egyptian

Really? Ancient Egyptian looks super Semitic to me. Do you think that's just an artefact of the reconstruction?

Well there is phonological similarity to Canaanite and Arabic's ancestors, but there isn't as much when you compare Semitic as a whole. There's also issues with lexical correspondence

mèþru wrote:Also, I'm not sure about the relationship between Semitic and Egyptian

Looking at some of the morphology such as the bound pronouns, it's difficult not to feel Egyptian and Semitic are related though.

Note: Egyptian ṯ is reconstructed as [tʃʰ].
1S -j (cf. Arabic -(n)ī)
2S.M -k (cf. Arabic -ka)
2S.F -ṯ (cf. Arabic -ki)
3S.M -f (cf. Arabic -hu)
3S.F -s (cf. Arabic -hā)
1D -nj
2D -ṯnj (cf. Arabic -kumā)
3D -snj (cf. Arabic -humā)
1P -n (cf. Arabic -nā)
2P -ṯn (cf. Arabic -kum and -kunna)
3P -sn (cf. Arabic -hum and -hunna)

The nominal morphology also had -w or -ww for the plural of masculine nouns, and -t, -jt or -wt for feminine nouns.

Alon wrote: ↑Thu Jan 17, 2019 5:59 am But split ergativity only evolved recently in Indo-Aryan, in the early to mid second millennium, not even reaching unrecognized languages that are considered dialects of Hindi, like Bhojpuri.

Like I said, my knowledge of II syntax is passing at best.

mèþru wrote: ↑Thu Jan 17, 2019 6:54 amAlso, I'm not sure about the relationship between Semitic and Egyptian

I've seen this idea tossed around by a couple users on this board before, but I've never seen the relationship questioned by specialists in the field. While greater Afroasiatic may be questionable, I think "Boreoafroasiatic" (Berber-Egyptian-Semitic) is rather solid.

Ser wrote: ↑Thu Jan 17, 2019 11:55 am

mèþru wrote:Also, I'm not sure about the relationship between Semitic and Egyptian

Looking at some of the morphology such as the bound pronouns, it's difficult not to feel Egyptian and Semitic are related though.

Agreed.

Alon wrote: ↑Thu Jan 17, 2019 5:59 am But split ergativity only evolved recently in Indo-Aryan, in the early to mid second millennium, not even reaching unrecognized languages that are considered dialects of Hindi, like Bhojpuri.

But one can see the early stage in Pali, and I believe the history of Sanskrit shows it gathering strength. One can also see it in the order of the cases - the first three are nominative, accusative and instrumental, and it is the instrumental that serves as the ergative.

Linguoboy wrote: ↑Thu Jan 17, 2019 9:46 am

mèþru wrote: ↑Thu Jan 17, 2019 6:54 amAre South Semitic languages mainly VSO either historically or currently?
Also, I'm not sure about the relationship between Semitic and Egyptian

Really? Ancient Egyptian looks super Semitic to me. Do you think that's just an artefact of the reconstruction?

There are some weird features, though, like the prefix gender/number markings p/t/n for masculine/feminine/plural. The inflexions mostly have a Semitic feel, but it does make one wonder if it might just be borrowing. And the lexical match does feel remarkably low.

The grammatical features may be due to areal influence.

mèþru wrote: ↑Tue Jan 22, 2019 1:22 pmThe grammatical features may be due to areal influence.

And the low number of lexical matches might be due to substratum effects.

I definitely am not sold on Afroasiatic as a whole, but agree that there's enough to strongly suggest that at least Semitic, Berber, and probably Egyptian are distantly related, with the person-marking, general morphological template, and nominal morphology being the main evidence. ("Grammatical features" like a templatic morphology could come to resemble each other through areal influence, but it's unlikely that "grammatical features" like the Berber and Semitic verbal personal affixes -- specifically t- for 3sf/2s/2p, y- for 3sm, n- for 1p -- could come to resemble each other so closely just through areal influence. [I'd say it's much more likely that it's due to chance than due to areal influence, if it's not due to genetic inheritance!])* But it's true that no one has actually successfully proved this yet (the existing Afroasiatic reconstructions have been mostly laughable and completely contradictory). And then there's the issue that there's two totally different reconstructions of early Egyptian phonology, which means the people looking for cognates with Semitic have turned up two sets of cognate lists with totally different members that contradict each other, which is ... not inspiring.

Anyway, this discussion was originally in the context of whether Egyptian's word order typology has any bearing on what we can reconstruct for PS word order and my response would be that even if they are related, the obvious answer is "no," because they're both (a) distantly related, assuming they are related, and one or the other could easily have changed its word order over thousands of years, and (b) Egyptian was spoken in the same region as many Semitic languages and they could easily have influenced each other syntactically (at least the languages of the Levant).


*For example, the normal Ghadames person/number/gender verbal affixes are:
1s Ø -- ăʕ [< *-ăɣ] (cf. PS Stative/SC *-ku ?)
2s t -- ət [< *-əd] (cf. PS PC 2sm *ta -- Ø, 2sf *ta -- ī; PS Stative/SC 2sm *-ti, 2sf *-ta ?)
3sm y -- Ø (cf. PS PC *ya -- Ø)
3sf t -- Ø (cf. PS PC *ta -- Ø)
1p n -- Ø (cf. PS PC *ni -- Ø)
2pm t -- ăm (cf. PS PC *ta -- ū)
2pf t -- măt (cf. PS PC *ta -- ā)
3pm Ø -- ăn
3pf Ø -- năt


Ninja edit:

Frislander wrote: ↑Thu Jan 17, 2019 8:41 am

mèþru wrote: ↑Thu Jan 17, 2019 6:54 am Are South Semitic languages mainly VSO either historically or currently?
Also, I'm not sure about the relationship between Semitic and Egyptian

Basically the only groups that are or have been VSO are Northwest Semitic and Arabic and its close relatives. Modern South Arabian is pretty universally SVO afaict, and Ethiosemitic universally SOV except for Ge‘ez, and there's reason to think they might have been intentionally mirroring the verb-initial syntax of its northern neighbours (in fact imho Ge‘ez suffers generally from people projecting onto it from a Proto-Semitic that had already been reconstructed based on Asian Semitic, which is just the wrong war round tbh). Additionally Ethiosemitic is a much deeper and more diverse branch of the family than many people realise, to the extent that it might actually be better to posit the PS Urheimat in the horn of Africa.

MSA languages apparently have VSO word order when the subject is an overt nominal/NP and SVO order when the subject is an overt pronoun (source: Marie-Claude Simeone-Senelle, "Modern South Arabian" in The Semitic Languages: An International Handbook, ed. Stefan Weninger et al, p. 1101). The basic order for Ancient South Arabian also seems to have been VSO, from what I can tell, except in the first clause of an inscription. There's also the fact that the Cushitic and other languages neighboring ES mostly have SOV word order, and since ES is part of a Sprachbund with those languages it could easily have adopted SOV order as a result of contact, as it did with some other features. It seems at least as reasonable to posit VSO as the Proto-Semitic order as to posit something else, since the languages which *don't* show VSO are primarily the ones that have had intensive contact with languages with different word orders.