What is the closest language to Arabic?

[k8an]

Yes, I'm fairly convinced that Arabic is closer to Hebrew than it is to Aramaic, but would you agree that Hebrew and Aramaic are still closer to each other than either of them is to Arabic?

Yeah, I think so - but it can be pretty hard to determine considering how divergent Aramaic dialects are. Old dialects of Aramaic (especially the Jewish ones) seem really similar to the point that it's almost mutually intelligible. Modern dialects are a different story though - centuries of Iranian, Turkic and Arabic influence on Assyrian/Eastern dialects have made them barely any more intelligible than Arabic at first listen (and some of the dialects from Nineveh in Iraq have been heavily influenced by Arabic vocabulary in the last 30-50 years as well). The dialects of Syria are even more heavily influenced by Arabic, and barely spoken enough for me to get a real idea. Of course, Modern Hebrew is also different to older incarnations of the language.

With all that said though, I remember listening to some Assyrian songs before I had ever studied the language and being able to pick out entire phrases (4-5 words) and understand them. I feel like the vocabulary and some aspects of word formation of Aramaic and Hebrew are closer than either is to Arabic, but Arabic and Hebrew have really similar verb conjugations which always trips me out.

It looks more like a Greek-inflected version of Maltese to me. But like I said (years ago), languages like Maltese, Cypriot Arabic and the even more exotic Arabic dialects of central Asia are all part of what we call "Arabic" so bringing them into this discussion doesn't make much sense.

That's a good point.

 

[Schem]

Here is him reading some Sabaic if you are interested.

If I were to use Islamic genealogical terminology, I would say this sounds like an old Qahtanite Arabic while ours is the Ishmaelite version influenced by the Semitic tongues of the Levant hence the association to Hebrew. Taxonomically, of course, they're all sister languages having inherited features independently and having influenced eachother through geography with northern Arabic dialects acting as a bridge inbetween westerly groups.

 

[radagasty]

Note that even Hebrew has broken plurals, but it consistently adds either -im or -ot to the plural e.g sefer -> sefarim على وزن فِعل التي تصبح فِعال.

I would say that Hebrew doesn't have broken plurals in the same sense as Arabic, where one has to know that the plural of صِفْر is أصفار, the vocalisation of which cannot be predicted (although there are recurring patterns). In Hebrew, however, the plural of segholate nouns have a fixed vocalisation /ə-ā/, irrespective of the root vowel in the singular, thus סְפָרִים → סֵ֫פֶר, following the model of the non-segholate דְּבָרִים → דָּבָר.

To take another example of a Hebrew segholate, אֲרָצוֹת → אֶ֫רֶץ, the cognate of which in Arabic is أَرْض, which has both a sound plural أَرَضُون and a broken plural أَرَاضٍ‎, the Hebrew plural corresponds to the former rather than the latter in its vocalisation. Thus, even though the pluralisation of סֵ֫פֶר and אֶ֫רֶץ does entail the appearance of a vowel /ā/ where there was none before (the seghol under the second radical being merely epenthetic, as its absence in their Arabic cognates suggests), I would say this phenomenon is more akin to the appearance of the fatha in the sound plural أَرَضُون, and therefore not an example of a broken plural, which أَرْض likewise has. That this vowel is long in Hebrew—it was historically short—is merely a result of a phonological rule that lengthens the vowel in an open pre-tonic syllable .

 

[Wadi Hanifa]

Akkadian has also the same case markers but it is in terms of grammar ( conjugation and almost all prepositions are different) and vocabulary very special.

Yes, but if we're trying to see which of NW Semitic and ASA is closer to Arabic, this is another point in favor of ASA. Now it could be that:

1) Arabic and ASA just happened to retain this case system (and other features like phonology and broken plurals) from Proto-Semitic independently of each other (so ASA, NW and Arabic are three independent branches of West Semitic), or

2) ASA and Arabic share these features because they share a more recent common ancestor that had them (so some sort of SW branch), or

3) Arabic is just a particularly conservative sibling of the NW languages, so that its similarities with ASA are just common Proto-Semitic retentions (like the Akkadian case system) (feels unlikely), or

4) Aramaic is a particularly innovative sibling of Hebrew and Arabic (in which case why are Hebrew and Aramaic grouped together as NW?)

Broken plurals as far as I understand are not an original thing of Semitic languages and mainly developed in the south of Arabia and in some Ethiopic languages as some kind of Sprachbund. Note that even Hebrew has broken plurals, but it consistently adds either -im or -ot to the plural e.g sefer -> sefarim على وزن فِعل التي تصبح فِعال. This phenomenon even exist in some Syrian words e.g. قرقوعة (scrap or useless thing) -> قراقيع qraaqee3 -> قراقعين qraaq3een

As @radagasty said, this Hebrew plural form is probably not a broken plural at all, and certainly can't be compared to the very extensive and well developed system of Arabic. Is the South Arabian (e.g. Sabean) plural system similar to the Arabic one? If it is, then that seems like pretty strong evidence for a closer relationship. The Ancient North Arabian languages also seemed to have broken plurals, which I think adds weight here.

By the way, the قراقعين example occurs in some Najdi dialects too (كبير > كبار > كبارين). Seems like just a 'plural of a plural' that arises independently and doesn't indicate any shared retention with Hebrew.

I highly recommend you to follow Ahmad Al-Jallad on Twitter if you have not already. Here is him reading some Sabaic if you are interested.

I do and I'm half-way through one of his papers on South Arabian languages. Does he tackle this question of classification/genetic descent directly anywhere?

If I were to use Islamic genealogical terminology, I would say this sounds like an old Qahtanite Arabic while ours is the Ishmaelite version influenced by the Semitic tongues of the Levant hence the association to Hebrew. Taxonomically, of course, they're all sister languages having inherited features independently and having influenced eachother through geography with northern Arabic dialects acting as a bridge inbetween westerly groups.

No please not the genealogical terminology!

This was the conflation that caused people like Taha Hussein to think that Imru' Al-Qays didn't speak Arabic so couldn't have composed Arabic poetry. Most of the tribes the genealogists called "Qahtanite" (including the real-life historical Qahtan) spoke North Arabian Arabic. Only a few were described as speaking "Himyarite" languages. The Muslim scholars never claimed that the Qahtanites spoke another language (they thought the Qahtanites taught Arabic to the Adnanites). It was only in modern times that early orientalists and their disciples like Taha Hussein made the logical leap that North Arabian vs. South Arabian languages is equivalent to Adnanite v. Qahtanite tribal genealogies, and so they thought Qahtanite = South Arabian and Qahtanite language = South Arabian language.

But yes I agree the simplest explanation is that North Arabian languages (like Arabic) are a sort of intermediate form between NW languages of the Levant like Canaanite/Hebrew and Ancient South Arabian languages like Sabaic. If I were to venture to guess, I would say there was an early West Semitic speaking group in the Levant that split off and migrated south and became the Ancient South Arabians, and a group that migrated south slightly later whose language became Arabic, and perhaps a third a group that stayed in the Levant (or returned there early on) whose language became some of the other Ancient North Arabian dialects (like Safaitic). And then areal influence pulled the Arabic inside the peninsula towards South Arabian and pulled the other Ancient North Arabian dialects towards Aramaic. You probably also had later Arabic tribal migrations from inside the Peninsula before Islam (like the Iyad, Tanukh, Lakhm, Ghassan, Kalb, Jutham, Taghlib, etc.) bringing the Arabic we know today and and continuing until modern times, where older groups and dialects are displaced by dialects from the interior.

 

[Ihsiin]

Yes, but if we're trying to see which of NW Semitic and ASA is closer to Arabic, this is another point in favor of ASA. Now it could be that:

A few points occurs to me...

1) Arabic and ASA just happened to retain this case system (and other features like phonology and broken plurals) from Proto-Semitic independently of each other (so ASA, NW and Arabic are three independent branches of West Semitic), or

2) ASA and Arabic share these features because they share a more recent common ancestor that had them (so some sort of SW branch), or

Well, I believe some ancient NW Semitic languages such as Ugaritic also have the case system, so it seems the case system was retain in proto-NW Semitic and lost in later stages of the languages (actually, just the same as in Arabic). Also, on the phonology front, I believe the way in which proto-Semitic phonemes merged across the NW Semitic languages implies that most if not all of the phonemes were conserved in photo-NW Semitic, so this doesn't really tell us much about its relationship to Arabic. Really the only feature I see that Arabic shares with Ancient South Arabian and not with NW Semitic is broken plurals.

3) Arabic is just a particularly conservative sibling of the NW languages, so that its similarities with ASA are just common Proto-Semitic retentions (like the Akkadian case system) (feels unlikely), or

Personally I find this to be the most likely scenario. As I mentioned, I think the only feature that needs to be 'explained' between Arabic and ASA is the broken plural system, and I can well believe this to be a proto-Semitic feature lost in daughter languages as they normalise towards regular plurals, and then only retained in the languages that happen to retain them. Even in Arabic, in some modern dialects we can see some instances of plural normalisation, such as حلمين instead of أحلام.

I also have this kind of idea that Semitic languages spoken by nomadic peoples tend to be more conservative than those spoken by sedentary peoples. Arabs of course remained predominantly nomadic for much longer than other Semitic peoples so perhaps it should not be surprising that Arabic retained many older Semitic features beyond the point when many of its sister languages lost them.

4) Aramaic is a particularly innovative sibling of Hebrew and Arabic (in which case why are Hebrew and Aramaic grouped together as NW?)

I think we need to be mindful of the stage of the language we're looking at. Remember Hebrew stopped being used as a spoken language around 2,000 years ago, so modern Aramaic has roughly 2,000 years of divergent evolution on Hebrew. Similarly, we continue to have knowledge of Classical Arabic which was spoken around ~1,500, and though there has been a lot of evolution within spoken Arabic since then, our knowledge of the classical language props up that particular familiarity with Hebrew. Perhaps if we only knew modern vernacular Arabic, Hebrew would not look so similar. For example, Hebrew מה means 'what', and this is instantly recognisable as being the same as Arabic ما - but this is in Classical Arabic, pretty much every modern Arabic dialect has a ش-type word to mean 'what', and ما is used exclusively for negation (which it isn't in Hebrew). So, the later the stage of the language we look at, the fewer similarities we find. I suspect (and perhaps some learned Aramaicists can assist us on this), that if we look at an ancient form of Aramaic rather than at Neo-Aramaic, we will find it to be much more similar to both Arabic and Hebrew.

 

[Wadi Hanifa]

A few points occurs to me...

Well, I believe some ancient NW Semitic languages such as Ugaritic also have the case system, so it seems the case system was retain in proto-NW Semitic and lost in later stages of the languages (actually, just the same as in Arabic). Also, on the phonology front, I believe the way in which proto-Semitic phonemes merged across the NW Semitic languages implies that most if not all of the phonemes were conserved in photo-NW Semitic, so this doesn't really tell us much about its relationship to Arabic. Really the only feature I see that Arabic shares with Ancient South Arabian and not with NW Semitic is broken plurals.

Yes but you're comparing apples and oranges (e.g. Old Arabic and ASA South Arabian from around 500-600 CE with NW languages almost 2000 years before). It may not be conclusive proof but it least suggests that North Arabian and South Arabian may have had a more recent common ancestor that split off from the NW languages.

Personally I find this to be the most likely scenario. As I mentioned, I think the only feature that needs to be 'explained' between Arabic and ASA is the broken plural system, and I can well believe this to be a proto-Semitic feature lost in daughter languages as they normalise towards regular plurals, and then only retained in the languages that happen to retain them. Even in Arabic, in some modern dialects we can see some instances of plural normalisation, such as حلمين instead of أحلام.

I'm going to have to strongly disagree with you here. The tendency in vernacular Arabic has certainly not been away from broken plurals. The system from Classical Arabic is more or less intact in modern dialects (marginal phenomena like حلمين notwithstanding) and if anything the broken plurals seem to be preferred. The influence of "MSA" in the modern period may have brought back some non-broken plurals into use (e.g. إيطاليين vs طليان), but this seems like hypercorrection. This is very different from the situation with the case-marking system. There is no evidence of a tendency for the broken plural system to 'decay' with time, and it seems to me that if NW languages had them at the time they separated from Arabic they would have retained them long enough for them to be attested.

Of course there is no dispute that Arabic, ASA and the NW languages all share a common ancestor, and there also seems to be a consensus that Hebrew and Aramaic form their own sub-branch (NW Semitic). I guess what I'm trying to say is I don't understand the reason why we have to come up with a "Central Semitic" group that includes Arabic and NW Semitic but excludes ASA. Seems like the situation is more likely to be this:

West Semitic splits into:

1) NW Semitic (splits into Canaanite and Aramaic, and Canaanite splits into Hebrew, Phoenician, Ugaritic, etc.)
2) Ancient North Arabian (including Arabic)
3) Ancient South Arabian (includes Sabaic, Himyaritic, etc.)

So branch 2 may not be particularly close to either 1 or 3, but if you want to decide which is most similar to 2 you have to look at common features regardless of how far back they go. Looking at things like broken plurals, phonology and case markings, it seems that Classical Arabic at least is more like ASA than Hebrew, and there is at least a case that actually 2 and 3 used to be a single language that broke off from the common ancestor of 1, 2 and 3.

I also have this kind of idea that Semitic languages spoken by nomadic peoples tend to be more conservative than those spoken by sedentary peoples. Arabs of course remained predominantly nomadic for much longer than other Semitic peoples so perhaps it should not be surprising that Arabic retained many older Semitic features beyond the point when many of its sister languages lost them.

Perhaps, though it probably has more to do with geographic isolation than lifestyle since the ASA speakers were sedentary as well. In any case, the modern vernaculars are also relatively conservative compared to NW languages (even compared to their older forms other than Ugaritic I guess).

I think we need to be mindful of the stage of the language we're looking at. Remember Hebrew stopped being used as a spoken language around 2,000 years ago, so modern Aramaic has roughly 2,000 years of divergent evolution on Hebrew. Similarly, we continue to have knowledge of Classical Arabic which was spoken around ~1,500, and though there has been a lot of evolution within spoken Arabic since then, our knowledge of the classical language props up that particular familiarity with Hebrew. Perhaps if we only knew modern vernacular Arabic, Hebrew would not look so similar. For example, Hebrew מה means 'what', and this is instantly recognisable as being the same as Arabic ما - but this is in Classical Arabic, pretty much every modern Arabic dialect has a ش-type word to mean 'what', and ما is used exclusively for negation (which it isn't in Hebrew). So, the later the stage of the language we look at, the fewer similarities we find. I suspect (and perhaps some learned Aramaicists can assist us on this), that if we look at an ancient form of Aramaic rather than at Neo-Aramaic, we will find it to be much more similar to both Arabic and Hebrew.

That's true, but what I've had in mind in this discussion was the state of the languages around the time when ASA and Old Arabic existed, at least for the purposes of figuring out how they all relate to each other genetically. I agree that modern Arabic dialects look a bit more like Aramaic than older ones, but not substantially so. (By the way Yemeni Arabic dialects routinely use ما as an interrogative, and most other Peninsular dialects still use it as a relative pronoun.)

 

[Ihsiin]

Yes but you're comparing apples and oranges (e.g. Old Arabic and ASA South Arabian from around 500-600 CE with NW languages almost 2000 years before). It may not be conclusive proof but it least suggests that North Arabian and South Arabian may have had a more recent common ancestor that split off from the NW languages.

Not quite. My point here was to note that if we reconstruct proto-NW Semitic we actually find a phonology and grammar very similar to Arabic, and I don't think less similar than ASA. The fact that subsequent NW Semitic languages diverged and lost some of these features doesn't really mean much when we look at the languages phylogenetically, as it were. When considering how the languages are related to each other (that is to say, from which common ancestors did they descend), we need to look at the oldest forms of the languages possible.

I'm going to have to strongly disagree with you here. The tendency in vernacular Arabic has certainly not been away from broken plurals. The system from Classical Arabic is more or less intact in modern dialects (marginal phenomena like حلمين notwithstanding) and if anything the broken plurals seem to be preferred. The influence of "MSA" in the modern period may have brought back some non-broken plurals into use (e.g. إيطاليين vs طليان), but this seems like hypercorrection. This is very different from the situation with the case-marking system. There is no evidence of a tendency for the broken plural system to 'decay' with time, and it seems to me that if NW languages had them at the time they separated from Arabic they would have retained them long enough for them to be attested.

I completely agree that broken plurals remain dominant in vernacular Arabic. My example of حلمين was really just to suggest how broken plurals could be replaced by regular plurals as some sort of paradigm levelling. All NW Semitic languages have regular plurals so if we were to suggest a shift from broken to regular plurals this would have been at a point before proto-NW Semitic split up into its daughter languages, so well before we would expect such a thing to be attested. I do admit, this is supposition and we don't have direct evidence for it, but it's a way of trying to rationalise the relative similarities between Arabic, NW Semitic and Ancient South Arabian.

(P.S. طليان means 'lambs' in Iraqi (and I guess other dialects too?), so perhaps ايطاليين has more utility in this case - made me chuckle, anyway).

Of course there is no dispute that Arabic, ASA and the NW languages all share a common ancestor, and there also seems to be a consensus that Hebrew and Aramaic form their own sub-branch (NW Semitic). I guess what I'm trying to say is I don't understand the reason why we have to come up with a "Central Semitic" group that includes Arabic and NW Semitic but excludes ASA. Seems like the situation is more likely to be this:

West Semitic splits into:

1) NW Semitic (splits into Canaanite and Aramaic, and Canaanite splits into Hebrew, Phoenician, Ugaritic, etc.)
2) Ancient North Arabian (including Arabic)
3) Ancient South Arabian (includes Sabaic, Himyaritic, etc.)

So branch 2 may not be particularly close to either 1 or 3, but if you want to decide which is most similar to 2 you have to look at common features regardless of how far back they go. Looking at things like broken plurals, phonology and case markings, it seems that Classical Arabic at least is more like ASA than Hebrew, and there is at least a case that actually 2 and 3 used to be a single language that broke off from the common ancestor of 1, 2 and 3.

Well, we also find broken plurals in Ge'ez, and I believe in a very reduced state in Amharic (I suppose this is also probably a better example of a shift from a broken plural system to a regular plural system, assuming broken plurals to be a feature in proto-Ethiopic). Ge'ez also shares some features with ASA not shared with Arabic, such as using /k/ to form suffixes in the perfect tense as opposed to /t/ used in Arabic and NW Semitic (e.g. فعلكُ vs فعلتُ). But when I look at Ge'ez more broadly it seems a lot less similar to Arabic than Hebrew does, for example. But then I suppose this is the point where I raise my hands and say that I am not a linguist and a feeling of how similar languages are is not the same as a robust and comprehensive comparison of linguistic features. I would say, however, that I don't see that Arabic has more in common with ASA than it does with NW Semitic.

Perhaps, though it probably has more to do with geographic isolation than lifestyle since the ASA speakers were sedentary as well. In any case, the modern vernaculars are also relatively conservative compared to NW languages (even compared to their older forms other than Ugaritic I guess).

Yeah, I don't know, it was just a pet theory I had, I really have nothing robust on which to base it

That's true, but what I've had in mind in this discussion was the state of the languages around the time when ASA and Old Arabic existed, at least for the purposes of figuring out how they all relate to each other genetically. I agree that modern Arabic dialects look a bit more like Aramaic than older ones, but not substantially so. (By the way Yemeni Arabic dialects routinely use ما as an interrogative, and most other Peninsular dialects still use it as a relative pronoun.)

Yeah, I was just suggesting that Aramaic seems to be so different from Hebrew and Arabic because we're looking at Neo-Aramaic, not ancient Aramaic.

 

[momai]

Yes, but if we're trying to see which of NW Semitic and ASA is closer to Arabic, this is another point in favor of ASA. Now it could be that:

1) Arabic and ASA just happened to retain this case system (and other features like phonology and broken plurals) from Proto-Semitic independently of each other (so ASA, NW and Arabic are three independent branches of West Semitic), or

2) it could be that ASA and Arabic share these features because they share a more recent common ancestor that had them (so some sort of SW branch), or

3) Arabic is just a particularly conservative sibling of the NW languages, so that its similarities with ASA are just common Proto-Semitic retentions (like the Akkadian case system) (feels unlikely), or

4) Aramaic is a particularly innovative sibling of Hebrew and Arabic (in which case why are Hebrew and Aramaic grouped together as NW?)

As I said earlier broken plurals seem from many papers I read to have developed later evident by the fact that not even Akkadian 2500 BC. used them and semitists don't seem to consider broken plurals as much important as we would like to think when they go and classify semitic languages. It is of course one aspect of Semitic languages but only one.
As for your question why Aramaic and Hebrew are grouped together, this of course can't be explained here in full detail, but some points off my head:
1- initial waw becomes y as in yeled -> walad , wasan -> yeshen (to sleep)
2- The realization of S123 in NWS
s1 -> Ar: s -> NW: sh
s2 -> Ar: sh -> NW: samekh
s3 -> Ar: s -> NW: s
So Arabic doesn't have the samekh letter
3- much more similar vocabulary

I do and I'm half-way through one of his papers on South Arabian languages. Does he tackle this question of classification/genetic descent directly anywhere?

I don't know any of him, but here are two papers discussing ASA on Academia Aljallad linked them on his Twitter
one two
I have a dictionary of Sabaic (and many other semitic languages) I downloaded sometime earlier I could send to you on private if you wish.
One of the paper discusses how ASA use a b- prefix to denotes the present/future tense which is exactly what we still have in dialects.
The other interesting aspect is the w- prefixed verbs which are used for the past tense which is exactly how it is found in Biblical Hebrew.
Both of these prefixes didn't exist in Classical Arabic so ....