Triconsonantal Root Systems

[Mecislau]

Interesting, although I believe a better cognate form in Arabic would be istafʕala, with the consonant cluster stoping the change /s/ > /h/. Actually, the whole /s/ > /h/ change is kind of strange, at least based on the corresponding pronouns & verb augments from East Semitic, PS *šalāmum should've become *halāmun and *halōm in Arabic and Hebrew, respectively.

(Well, I didn't actually mean those two forms were the closest - I'm just saying hitpa`el does the same thing)

And while I'm clearly no expert, I'm more apt to think that it was the change to /h/ that was irregular, since you only see it in a few different forms (that I'm aware of).

I think one of the things that make comparative Semitic studies more difficult than, say, comparative IE is that most Semitic languages were in mutual contact for several thousand years; it makes it very difficult to distinguish between inherited features/shared innovation and borrowings.

Well, it's not like IE languages weren't in mutual contact for several thousand years either...

Out of curiosity, do you use the root T-R-G-M in Hebrew to mean translate? and if so, which class does it belong to?

Yes, and it's a regular quad-consonantal pi`el verb.

So /yeš/ means "there is" in Hebrew? That's very interesting, because part of me is wondering if יש is related to ليس /laysa/ in Arabic - that is Arabs used the construction *lā + *yas ("no" + "there is") for "there is not" so much that they fused together. It certainly would explain why it behaves like no other verb in the language.

In fact, I have seen that very explanation before.

A paper I was just reading explains that yesh (as well as the Arabic equivalent) was originally an emphatic particle, composed of a prefix *ye- (as seen in the third person future/imperfective of verbs) and a demonstrative *š, so a phrase such as "yesh li X" (modern Hebrew "I have X", lit. "there is X to me") would have originally meant more of a "Behold X is to me". This form is apparently seen in Egyptian as *ys and Coptic as eys, where it is used as "behold".

Just a random tidbit.

The word עוד looks like it might be cognate of عاد /ʕāda/ "he returned". Amongst other things, when used as an auxiliary it means something like "he did it again", especially in the negative: ما عاد يحاربون /mā ʕāda yataħārabūna/ "they didn't fight again" (lit. not return they-fight), لم يعد السفرة ممكنةً /lam yaʕudi-s-safratu mumkinatan/ "the trip wasn't possible any more" (lit. not returned-the-journey (as) possible).

I wouldn't be surprised if they were related. In addition to meaning "still", עוד also means "another" (עוד פעם `od pa`am "another time, again")

Two questions: Is /ke'an/ spelled with 'ק' or 'כ'? for that matter was the glottal stop 'א' or 'ע'?

כאן

Although you probably haven't noticed, I almost always transcribe ק as "q" and כּ as "k", and א as ' and ע as `.

Also, could you give me an example of הרי in a sentence?

Sure. Here's another googled example (like I said, it's rather legalistic):

הריני מתחייב לשלם את מלוא שכר הלימוד עבור הקורסים אליהם נרשמתי
Hareyni mitchayev leshalem 'et milu' shachar halimud `avur haqursim 'eleyhem nirshamti
"I hereby promise to pay the entirety of the tuition for the courses for which I have registered."

I'm wondering if bilti is related to Arabic بل /bal/, which means something like, "but rather" or "on the contrary": لا أتكلّم العبرية بل العربية /lā 'atakallamu-l-ʕibriyyata bal 'al-ʕarabiyyata/ "I don't Hebrew, but (in contrast/rather) Arabic", لا أدرس اللغات الحديثة فحسب بل والقديمة أيضاً /lā 'adrusu-l-luɣāta-l-ħadīθata faħasbu bal wa-l-qadīmata 'ayḍan/ "I don't just study modern languages, but rather ancient ones as well".

Likely. Bilti comes from the negative form בל bal.

(Yes, more negative forms... Just yesterday I came across a list of different words use for negation throughout the Old Testament - the list came to nineteen different forms)

Of course, this is assuming the /t/ comes from 'ת' and not 'ט'. Curse you Ashkenazim for not being able to velarize/glottalize your consonants and thus making comparative Semitic etymology more difficult for me! *shakes fist*

Hah, I'll use the Hebrew script alongside the Latin from now on if you prefer (as I have yet to find a good way of distinguishing ת and ט in transcription without diacritics).

Yeah, it's בלתי.

[Wiseblood]

Sorry about taking so long to reply - RealLifeTM has kept me away.

(Well, I didn't actually mean those two forms were the closest - I'm just saying hitpa`el does the same thing)

Fair enough. It's actually pretty interesting how etymologically different forms can take on similar semantic roles. Actually, istafʕala verbs in Arabic often connotate "searching" or "asking": علم /ʕalima/ "he knew" > استعلم /istaʕlama/ "he inquired". It can also mean to "judge" or "percieve as": قبح /qabuħa/ "he was ugly" > استقبح /istaqbaħa/ "he loathed" (thought of as ugly).

Do you have a form in Hebrew that corresponds to Arabic Form V, tafaʕʕala? While tafaʕʕala verbs can also mean "to pretend" or "attempt to take the characteristics of" (ie: كبر /kabura/ "he was big" > تكبّر /takabbara/ "he self-aggrandized"), it can also mean something like "ethnicize". This is especiaaly true for what we call نسبة /nisbatun/ or nisbah words in Arabic (those that end in the derivational suffix -iyy):

x عرب /ʕarabun/ "Arabs" > عربيّ /ʕarabiyyun/ "Arab, Arabic" > تعرّب /taʕarraba/ "he arabacized" (himself)
x أمريكا /'amrīkā/ "America" > أمريكيّ /'amrīkiyyun/ "American" > تأمرك /ta'amraka/ "he americanized" (himself)
x بوذيّة /būðiyyatun/ "Buddhism" > تبوّذ /tabawwaða/ "he converted to Buddhism"

And while I'm clearly no expert, I'm more apt to think that it was the change to /h/ that was irregular, since you only see it in a few different forms (that I'm aware of).

Hmmm, if I'm understanding correctly, you're saying that the change /s/ > /h/ only occured in certain words and not in others. This makes perfect sense. Now, this may be a case of "six of one, half a dozen of the other", but I'd argue that it was the need of the proto-arabs & proto-hebrews to keep consonants consistant across semantically related words that stopped the change form occuring in words like *salāmum. Isn't anology awesome?

Well, it's not like IE languages weren't in mutual contact for several thousand years either...

Ouch! Touché!

Granted, several IE branches have historically been in contact (and still are) for several thousand years. My point, however, was that with IE, we can look at geographically seperated branches like Celtic, Aryan, and Tocharian, find common features and say, "well, that most certainly is a shared trait and not borrowing because these people haven't been in contact for ~3000 years". Semitic is somewhat murkier, because it was spread across a relatively smaller area. Now, if we were to find a common feature in Aramaic, Akkadian and Ge'ez, it would be difficult to say with the same amount of certainty if it was borrowed or not, as speakers of these languages were in contact for much of their (pre)history. The only really isolated branches that I'm aware of are the Modern South Arabian languages and the Southern Periphery Ethiopian languages, both of which belong to the same branch of Semitic anyway!

You are right, however - I am an idiot for making such a sweeping generalization.

In fact, I have seen that very explanation before.

A paper I was just reading explains that yesh (as well as the Arabic equivalent) was originally an emphatic particle, composed of a prefix *ye- (as seen in the third person future/imperfective of verbs) and a demonstrative *š, so a phrase such as "yesh li X" (modern Hebrew "I have X", lit. "there is X to me") would have originally meant more of a "Behold X is to me". This form is apparently seen in Egyptian as *ys and Coptic as eys, where it is used as "behold".

I'm guessing by *ys you mean Middle Egyptian jsw? *š/*sw being related to, say, Akkadian šū I can almost buy (altough Egyptian demonstratives don't look a whole lot like Semitic ones to me - they seem a lot more like Berber's noun prefixes). However, *ye- corresponding to *j- is a little bit more problematic. So far as I know, there is no trace of the Afro-Asiatic "prefix conjugation" in the Egyptian branch (the Old and Middle Egyptian verbal systems seem to be at least partially based on the AA "stative" or "suffix" conjugation). Also, I'm not aware (off the top of my head, granted) of any case where Semitic *y- corresponds to Egyptian *j-. Actually, Egyptian jmnt "west" may be related to Semitic *yamīn "right, right hand". However, j- often corresponds to Semitic *l-: jb vs. *libbum "heart", or *ʔ-: jnk vs. *'anāku (Coptic anok, I think) "I, me". Still, there are places where we'd expect to see a semi-vowel in Egyptian where there isn't one: compare egyptian ᶜ (this should be a superscript "c") "hand" versus PS *yadum (Egyptian ᶜ, [ʕ] often corresponds to Semitic /d/,/ð/ and /z/).

In truth, we still don't know a lot about Old Egyptian and Middle Egyptian phonology - especially in regards to how well hieroglyphic writing actually represented it. Still, I'd very much like to read that paper - could you tell me where I might find a copy?

I wouldn't be surprised if they were related. In addition to meaning "still", עוד also means "another" (עוד פעם `od pa`am "another time, again")

Very interesting. Are you aware of a Hebrew cognate of the Arabic word for another, آخر /'āxarun/?

Although you probably haven't noticed, I almost always transcribe ק as "q" and כּ as "k", and א as ' and ע as `.

Thanks, I'll try to pay more attention in the future. I asked because I suspect כאן is a cognate of Arabic كائن /kā'inun/ "being".

הריני מתחייב לשלם את מלוא שכר הלימוד עבור הקורסים אליהם נרשמתי

Okay, I have to admit I cracked up when I saw "הקורסים"! I take it that it's common in modern Hebrew to use "ק" instead of "כ" in borrowings to avoid lenition? Speaking of borrowings, I remember reading somewhere that "discuss" was borrowed in MH as well, is this true?

Hah, I'll use the Hebrew script alongside the Latin from now on if you prefer (as I have yet to find a good way of distinguishing ת and ט in transcription without diacritics).

Thanks, I'd appreciate it. I have to admit that while I'm fairly comfortable reading Hebrew consonants, all those vowel diacritics still scare the hell out of me!

As for transliteration, the two most common schemes I've seen online are letter + period (t.) and capitals (T). Granted, both schemes present problems (in particular, the capitals system brings back painful memories of flipping through the Klingon-English dictionary in a bookstore and almost going blind).

[hwhatting]

The only really isolated branches that I'm aware of are the Modern South Arabian languages and the Southern Periphery Ethiopian languages, both of which belong to the same branch of Semitic anyway!

Wouldn't the Southern Arabic languages have been in constant contact with Arabic?

[Mecislau]

Sorry about taking so long to reply - RealLifeTM has kept me away.

That's no problem. I understand.

Fair enough. It's actually pretty interesting how etymologically different forms can take on similar semantic roles. Actually, istafʕala verbs in Arabic often connotate "searching" or "asking": علم /ʕalima/ "he knew" > استعلم /istaʕlama/ "he inquired". It can also mean to "judge" or "percieve as": قبح /qabuħa/ "he was ugly" > استقبح /istaqbaħa/ "he loathed" (thought of as ugly).

Hmm, well, I don't think Hebrew really has an equivalent to those senses.

I'm guessing that etymologically-speaking the closest equivalent to Arabic istafʕala that Hebrew has is the minor binyan hishtaf`el. I can only think of one verb in it right off hand, though: השתחווה hishtachavah "bow down, prostrate".

That or hishtaf`el is just hitpa`el with an added Sh tagged to the beginning of the root. I really need to find out where these minor binyanim originate...

Do you have a form in Hebrew that corresponds to Arabic Form V, tafaʕʕala? While tafaʕʕala verbs can also mean "to pretend" or "attempt to take the characteristics of" (ie: كبر /kabura/ "he was big" > تكبّر /takabbara/ "he self-aggrandized"), it can also mean something like "ethnicize". This is especiaaly true for what we call نسبة /nisbatun/ or nisbah words in Arabic (those that end in the derivational suffix -iyy):

x عرب /ʕarabun/ "Arabs" > عربيّ /ʕarabiyyun/ "Arab, Arabic" > تعرّب /taʕarraba/ "he arabacized" (himself)
x أمريكا /'amrīkā/ "America" > أمريكيّ /'amrīkiyyun/ "American" > تأمرك /ta'amraka/ "he americanized" (himself)
x بوذيّة /būðiyyatun/ "Buddhism" > تبوّذ /tabawwaða/ "he converted to Buddhism"

I do not believe so. Hebrew seems to use hitpa`el, pi`el, or periphrastic constructions for this:

גדל gadal "he grew" > התגדל hitgadel "he praised himself, boasted" (~self-aggrandized)
נוצרי notzri "Christian" > ניצר nitzer "he christianized"
איסלאם 'islam "Islam" > איסלם 'islem "he islamicized"

Ethnicities and religious conversions use other verbs:
אמריקה 'ameriqah "American" > גרם לאמריקניזציה garam le'ameriqanizatzya (lit. "brought about Americanization" - the -atzya suffix clearly a Russian loan)
בודהיזם budhizm "Buddhism" > המיד (את דתו) לבודהיזם hemid ('et dato) lebudhizm (lit. "converted [his faith] to Buddhism").

I'm guessing by *ys you mean Middle Egyptian jsw? *š/*sw being related to, say, Akkadian šū I can almost buy (altough Egyptian demonstratives don't look a whole lot like Semitic ones to me - they seem a lot more like Berber's noun prefixes). However, *ye- corresponding to *j- is a little bit more problematic. So far as I know, there is no trace of the Afro-Asiatic "prefix conjugation" in the Egyptian branch (the Old and Middle Egyptian verbal systems seem to be at least partially based on the AA "stative" or "suffix" conjugation). Also, I'm not aware (off the top of my head, granted) of any case where Semitic *y- corresponds to Egyptian *j-. Actually, Egyptian jmnt "west" may be related to Semitic *yamīn "right, right hand". However, j- often corresponds to Semitic *l-: jb vs. *libbum "heart", or *ʔ-: jnk vs. *'anāku (Coptic anok, I think) "I, me". Still, there are places where we'd expect to see a semi-vowel in Egyptian where there isn't one: compare egyptian ᶜ (this should be a superscript "c") "hand" versus PS *yadum (Egyptian ᶜ, [ʕ] often corresponds to Semitic /d/,/ð/ and /z/).

In truth, we still don't know a lot about Old Egyptian and Middle Egyptian phonology - especially in regards to how well hieroglyphic writing actually represented it. Still, I'd very much like to read that paper - could you tell me where I might find a copy?

Well, I know very little about Egyptian, I'm just going by what the paper said, which talked about an Egyptian/Coptic particle *ys/*is and a suffixed form *yst/*ist (Coptic eys and este).

The paper was just something I was reading on the JSTOR database. If you have access, the paper is called "Hebrew and Semitic Particles" (one of a series of seven or so).

Very interesting. Are you aware of a Hebrew cognate of the Arabic word for another, آخر /'āxarun/?

Well, there's אחר 'acher "other" and אחרון 'acharon "last (in a series)".

Okay, I have to admit I cracked up when I saw "הקורסים"! I take it that it's common in modern Hebrew to use "ק" instead of "כ" in borrowings to avoid lenition?

Heh, yeah. They're pronounced the same in modern Hebrew, so ק is always used to represent /k/ in foreign loans because it completely avoids the issue of lenition.

Similarly, ט (equivalent to Arabic ط) is always used for /t/. This one is a little stranger, perhaps, since ת never lenites in modern Hebrew, but...

However, when borrowing from Arabic the most etymologically appropriate letter is used, so ת is used for ت and ט for ط . Hebrew ת is also used for /t/ derived from a foreign /T/, so you have words like תאטרון te'atron "theatre" with both types of T.

When it comes to foreign ב /b/, well, you just have to assume it doesn't lenite if it's a foreign word.

Interestingly, though, and I have no idea when this came about, but Hebrew does have an interesting method of marking foreign /p/ that does not lenite, at least when word-finally. Normally the letter פ (pey) takes its final form ף at the end of a word, where in native words it always lenites to [f]. However, in foreign words with final /p/, pey does not take its final form: סירופ sirop "syrup".

Speaking of borrowings, I remember reading somewhere that "discuss" was borrowed in MH as well, is this true?

Well, colloquially. The verb דסקס disqes is used in speech, but not in formal usage.


In other news, I'm excited to say that I'm finally going to take up Arabic (MSA for now). I've got a textbook and hopefully I'll be able to get into a basic MSA class this fall.

You wouldn't by any chance happen to know of a good online dictionary that includes vowels, would you? I've found a bunch of vowelless ones, but at this point that really doesn't help me a lot.

Seeing all the cognates in the basic vocabulary is quite interesting, though!

[Wiseblood]

Wouldn't the Southern Arabic languages have been in constant contact with Arabic?

Yes and no.

Firstly, the South Arabian languages aren't Arabic at all, but rather are more closely related to the Ethiopian Semitic languages (Ge'ez, Amharic, Chaha). I could go into a huge tangent about the classification of Semitic languages and how Arabic and the MSA languages fit in. I think, however, in place of boring you to tears, I'll just refer you to Wikipedia which, at least from my cursory glance over the articles, does a pretty good job of explaining the situation. The MSA languages are called "Eastern South Semitic" there.

We do know for a fact that Arabic speakers had contact with speakers of various Old South Arabian languages. However, MSA languages are not descended from them, but rather share a common ancestor. We can say this with some certainty because the MSA languages don't fully reflect the /s/ > /h/ sound change in pronouns and verbal augments that Maknas and I have been discussing. So far as I know, all Western Semitic languages went through this change, except for Qatabanic, an OSA language which coincidentally, happened to be spoken in an area adjacent to where many of the MSA languages are spoken today. Again, I'm trying really hard not to tangent here.

As for isolation, we can make arguments about whether or not the mainland MSA languages may have been "protected" from Semitic sprachbund by the OSA languages. However, I don't think anyone could make the case that the island of Soqotra wasn't isolated, at least historically.

I'm not certain if that was clear as I wanted it to be. Let me know if I actually answered your question.

I'm guessing that etymologically-speaking the closest equivalent to Arabic istafʕala that Hebrew has is the minor binyan hishtaf`el. I can only think of one verb in it right off hand, though: השתחווה hishtachavah "bow down, prostrate".

I'm actually fairly certain that the Hebrew cognate of istafʕala is hitpa`el, both reflecting PS *šatapaʕala. The intial /i/ is merely an epenthetic vowel to break up the consonant cluster at the beginning of a phrase, and doesn't represent the change /h/ > /ʔ/ found in 'afʕala verbs.

That or hishtaf`el is just hitpa`el with an added Sh tagged to the beginning of the root. I really need to find out where these minor binyanim originate...

I think this is probably the correct analysis. I wonder if there has ever been a study of these "minor binyanim"?

אמריקה 'ameriqah "American" > גרם לאמריקניזציה garam le'ameriqanizatzya (lit. "brought about Americanization" - the -atzya suffix clearly a Russian loan)

The derivational borrowing is striking. Are there many more examples of MH borrowing morphology from European languages?

As a side note جرم إلى /jarama 'ilā/ in Arabic means "he commited a crime against".

Well, I know very little about Egyptian, I'm just going by what the paper said, which talked about an Egyptian/Coptic particle *ys/*is and a suffixed form *yst/*ist (Coptic eys and este).

The paper was just something I was reading on the JSTOR database. If you have access, the paper is called "Hebrew and Semitic Particles" (one of a series of seven or so).

Fair enough. I'll have to find time to check it out, thanks.

Well, there's אחר 'acher "other" and אחרון 'acharon "last (in a series)".

Heh, compare Arabic آخَر /'āxarun/ "another" and آخِر /'āxirun/ "last".

Heh, yeah. They're pronounced the same in modern Hebrew, so ק is always used to represent /k/ in foreign loans because it completely avoids the issue of lenition.

Similarly, ט (equivalent to Arabic ط) is always used for /t/. This one is a little stranger, perhaps, since ת never lenites in modern Hebrew, but...

In Arabic, we tend to use whichever letters will most accurately reflect the original word's vowels. So, democracy is rendered as ديمقراطية /dīmaqrāṭiyyatun/ [ di:mɔqrˠɑ:tˠʷi:jaton ], but camera is rendered كاميرا /kāmīrā/ [ kɛ:mi:ra: ]. It can however, sometimes be rather arbitrary; compare أمريكا /'amrīkā/ [ ʔæmri:ka: ] "America" with موسيقى /mūsīqā/ [ mu:si:qa: ].

As a side note, IPA really sucks for making concise transcriptions of Arabic emphatics - [ tˠ ] doesn't really reflect ط very accurately. However, writing [ tˠˤʷ ] isn't much fun either.

However, when borrowing from Arabic the most etymologically appropriate letter is used, so ת is used for ت and ט for ط . Hebrew ת is also used for /t/ derived from a foreign /T/, so you have words like תאטרון te'atron "theatre" with both types of T.

We tend to do the same; for example, Shimon is rendered شمعون /šimʕūnu/. Actually, Hebrew borrowings in Arabic are confusing because some names reflect borrowings from Hebrew at an early stage (compare إسحاق /'isħāqu/ "Isaac" vs. native Arabic يضحك /yaḍħaku/ "he laughs"), while others may reflect shared Arabic traditions of Biblical stories ( قابيل وهابيل /qābīlu wa-hābīlu/ "Cain and Abel"). Still others merely reflect different pronunciations throughout history (modern إيهود /'īhūdu/ versus medieval يهوذ /yahūðu/).

I have to admit I've never quite understood why MH merges ת and ט. For some reason I find it hard to believe that Eastern Europeans couldn't velerize their consonants!

When it comes to foreign ב /b/, well, you just have to assume it doesn't lenite if it's a foreign word.

I can only assume that in Hebrew, much as it is in Arabic, that foreign words stick out so much that it's easy to tell the difference.

Interestingly, though, and I have no idea when this came about, but Hebrew does have an interesting method of marking foreign /p/ that does not lenite, at least when word-finally. Normally the letter פ (pey) takes its final form ף at the end of a word, where in native words it always lenites to [f]. However, in foreign words with final /p/, pey does not take its final form: סירופ sirop "syrup".

I have to say that that's a pretty cool solution. I guess it wouldn't be too much of a stretch to see these two developing into seperate letters at some point.

In other news, I'm excited to say that I'm finally going to take up Arabic (MSA for now). I've got a textbook and hopefully I'll be able to get into a basic MSA class this fall.

Good luck! Out of curiosity, what textbook is it?

Also, one piece of advice I'd like to give: when you start learning noun declination, try not to fall into the trap of thinking of the case European grammarians refer to as the "accusative" as being directly analogous to say, the Latin or German accusative. I've seen people (very bright, polyglots at that) get tripped up by this before.

You wouldn't by any chance happen to know of a good online dictionary that includes vowels, would you? I've found a bunch of vowelless ones, but at this point that really doesn't help me a lot.

Hmmm... the only one I'm aware of is the Sakhr Dictionary. However, it isn't terribly reliable with respect to uptime (as a matter of fact, as of this post, it has been down for two or three days).

I don't know if this is true for Hebrew, but because long vowels are always written in Arabic, I find I can predict a word's vowelling (even if I've never seen it before) just based on its "shape". This of course takes a little practice, and probably isn't much help to you now, but I though I'd mention it.

Seeing all the cognates in the basic vocabulary is quite interesting, though!

Tell me about it! Actually, I've been meaning to go through a "Modern Hebrew in 40 Lessons" book I have here. I think the only thing stopping me is the fact it's in French, and the transliteration scheme is designed for French speakers (the French do many things very well - designing orthographies isn't one of them).


A question: Are there compound tenses in Hebrew? For example, in Arabic we render the imperfective by way of "to be" in the perfect + the main verb in the imperfect: كان يكتب /kāna yaktubu/ "he used to write". Do you do something similar in Hebrew? If not, how would you express the imperfective, or (for example) the future perfect?

[hwhatting]

Wouldn't the Southern Arabic languages have been in constant contact with Arabic?

Yes and no.

Firstly, the South Arabian languages aren't Arabic at all, but rather are more closely related to the Ethiopian Semitic languages (Ge'ez, Amharic, Chaha).

Ana barif.
I was not talking about them being from the same branch, I was talking about them being geographical neighbours. You tackle that:

We do know for a fact that Arabic speakers had contact with speakers of various Old South Arabian languages. However, MSA languages are not descended from them, but rather share a common ancestor. We can say this with some certainty because the MSA languages don't fully reflect the /s/ > /h/ sound change in pronouns and verbal augments that Maknas and I have been discussing. So far as I know, all Western Semitic languages went through this change, except for Qatabanic, an OSA language which coincidentally, happened to be spoken in an area adjacent to where many of the MSA languages are spoken today. Again, I'm trying really hard not to tangent here.
As for isolation, we can make arguments about whether or not the mainland MSA languages may have been "protected" from Semitic sprachbund by the OSA languages. However, I don't think anyone could make the case that the island of Soqotra wasn't isolated, at least historically.

The MSA languages must have been in contact with Arabic dialects for all of their existence, and if anything has ever been between them and Arabic, it would have been other South Arabic languages - in any case, some form of Semitic. So there probably always (well, at least as far back as we're able to say anything about the linguistic situation there at all) was a Semitic continuum in Southern Arabia.
Even Soqotri would have been in contact with other SA languages and (later) with Arabic through sea trade.

So, whatever influence there actally was, what you said about the situation in regards to the Northern Semitic languages and Eastern Semitic would apply here as well - any communality needs to be double-checked on whether it's inherited or a case of loaning or an areal development. That's all the point I wanted to make.

[Wiseblood]

Ana barif.

Inta bta3raf? Mumtaz!

I was not talking about them being from the same branch, I was talking about them being geographical neighbours. You tackle that:

I have to admit that as I was typing all that, I thought to myself, "am I insulting his intelligence?". Clearly I was, my apologies. In all fairness, what threw me was "Southern Arabic languages". At least in a scholarly sense, "Southern Arabic languages" usually refer to Arabic as we know it today in contrast with "Northern Arabic languages" that pre-date Islam like, say Lihyanite. Basically, Arabic means one thing and Arabian means something else.

The MSA languages must have been in contact with Arabic dialects for all of their existence, and if anything has ever been between them and Arabic, it would have been other South Arabic languages - in any case, some form of Semitic. So there probably always (well, at least as far back as we're able to say anything about the linguistic situation there at all) was a Semitic continuum in Southern Arabia.

I think I should have said "relatively isolated" and not "really isolated". I would argue, however, that all evidence points to the MSA languages existing on the periphery of that continuum. It is probably circular logic, but the fact that the MSA languages didn't go through many of the changes that the "Great West Semitic SprachbundTM" did would tend to point to relative isolation (probably moreso cultural rather than physical) in my mind.

I'd also argue that we don't have a very clear linguistic picture of South Arabia prior to the early Common Era, long after South Semitic broke up. Before you say it - yes, I do realize the absence of proof doesn't equal proof of the opposite.

As for "must have" and "all of their existence", are you sure? For all we know, there were people speaking something related to Basque in South Arabia at some point.

Even Soqotri would have been in contact with other SA languages and (later) with Arabic through sea trade.

Good point. Granted, contact doesn't necessarily equal influence.

So, whatever influence there actally was, what you said about the situation in regards to the Northern Semitic languages and Eastern Semitic would apply here as well - any communality needs to be double-checked on whether it's inherited or a case of loaning or an areal development. That's all the point I wanted to make.

Like I said, the picture is pretty murky.

[Mecislau]

The derivational borrowing is striking. Are there many more examples of MH borrowing morphology from European languages?

Not that I know of. That's productive at least - I used the example of בודהיזם budhizm earlier, but that's borrowing the word, not the morphology.

You do see some Latin/Greek prefixes, however. They've been adapted to fit Hebrew syntax and morphology, though, since they appear as independent adverbs rather than as affixes: אנטי גלובליזציה 'anti globalizatzya. Of course, native equivalents exist too - the foreign adverbial "prefixes" seem to be used most often in technical jargon.

I have to admit I've never quite understood why MH merges ת and ט. For some reason I find it hard to believe that Eastern Europeans couldn't velerize their consonants!

Hah, I have enough trouble pronouncing [ tˠ ]. When you mentioned [ tˠˤʷ ], I nearly died...

I can only assume that in Hebrew, much as it is in Arabic, that foreign words stick out so much that it's easy to tell the difference.

Well, true. They tend to have a lot more consonants than the average Hebrew word and a lot more vowels being marked. It's quite hard to miss them.

Good luck! Out of curiosity, what textbook is it?

Well, it's the only decent looking one I found at the bookstore. It's called "Standard Arabic: An elementary-intermediate course". I got it for two dollars thanks to a bunch of giftcards I had never used...

Also, one piece of advice I'd like to give: when you start learning noun declination, try not to fall into the trap of thinking of the case European grammarians refer to as the "accusative" as being directly analogous to say, the Latin or German accusative. I've seen people (very bright, polyglots at that) get tripped up by this before.

Could you (briefly) explain what you mean?

I don't know if this is true for Hebrew, but because long vowels are always written in Arabic, I find I can predict a word's vowelling (even if I've never seen it before) just based on its "shape". This of course takes a little practice, and probably isn't much help to you now, but I though I'd mention it.

Well, naturally. I can predict the vowels in most Hebrew words fairly easily, but a lot of that comes from knowing the morphology. As for Arabic, well, as you said the long vowels aren't much of a problem - but I don't have much of an idea about the short vowels yet!

Tell me about it! Actually, I've been meaning to go through a "Modern Hebrew in 40 Lessons" book I have here. I think the only thing stopping me is the fact it's in French, and the transliteration scheme is designed for French speakers (the French do many things very well - designing orthographies isn't one of them).

Heh, sounds like fun.

A question: Are there compound tenses in Hebrew? For example, in Arabic we render the imperfective by way of "to be" in the perfect + the main verb in the imperfect: كان يكتب /kāna yaktubu/ "he used to write". Do you do something similar in Hebrew? If not, how would you express the imperfective, or (for example) the future perfect?

Well, there's one. It's formed with the past tense of "to be" ( = perfect) plus the present participles. It expresses three things:

Iterativity ("used to"):
היה כותב hayah kotev "he used to write"

Past or Present Conditionals ("would [have]")
אילו היה לו עט, היה כותב 'ilu hayah lo `et, hayah kotev "if he had a pen, he would have written"

Polite requests:
היית רוצה ללכת איתי? hayit rotzah lalechet 'iti? "Would you like to come with me?"


As for other tenses, aspects are just implied. Prepositional phrases can be employed if necessary: אני אכתוב את המכתב לא יאוחר מיום שישי 'ani 'echtov 'et hamichtav lo' ye'uchar miyom shishi "I will have written the letter by (lit. "no later than") Friday"

Now, if you don't mind, a couple of questions for you:

Do you know where the Arabic past tense forms of "be" like kāna are from? That form looks very different from the Hebrew hayah.

Why is wa- ("and" - I'll install an Arabic keyboard eventually) prefixed? Does it have any affect on words around it, or do other words affected it? In Hebrew at least ve- has a number of different forms depending on the structure of whatever it cliticizes to, and in some cases can affect the word itself (eg, ve + Yerushalyim "Jerusalem" > virushalayim). At least from what I've seen so far wa- doesn't look like it changes.

Historically-speaking, what does the dotless yaa signify? I realize it's /a:/ in MSA, but why is it used?

[Wiseblood]

Not that I know of. That's productive at least - I used the example of בודהיזם budhizm earlier, but that's borrowing the word, not the morphology.

This got me to thinking about how we tend to naturalize foreign words in Arabic. As an example, the Classical Arabic word for "orange" (the fruit) is برتقال /burtaqālu/ (clearly a greek loan). This is decidedly "un-semitic". Over time the word changed to become more in line with other noun patterns:

burtaqālu
*burtaqāl (loss of case ending)
*burta'āl ( /q/ > /ʔ/)
*burt'āl (unstressed vowel is dropped)
*burṭāl (glottal stop is dropped to avoid consonant cluster, /t/ becomes emphatic due to glottal influence)
*burṭān (/l/ > /n/)

The final form, burṭān looks like any other word of the fuʕlān type (bustān "garden", burdān "Cold").

You do see some Latin/Greek prefixes, however. They've been adapted to fit Hebrew syntax and morphology, though, since they appear as independent adverbs rather than as affixes: אנטי גלובליזציה 'anti globalizatzya. Of course, native equivalents exist too - the foreign adverbial "prefixes" seem to be used most often in technical jargon.

Interesting, my understanding is that Maltese does similar things. I think it'd be rather interesting to do a study comparing European influences on both Modern Hebrew and Maltese.

For the record, we tend to use native vocabulary for such purposes in Arabic: ضدّ العولمة /ḍidda-l-ʕawlamati/ "anti globalization" (lit. against-the-worlding).

Hah, I have enough trouble pronouncing [ tˠ ]. When you mentioned [ tˠˤʷ ], I nearly died...

It's not as hard as you think (at least if you can make a "dark" 'l' in English). Say 'mill' and hold your tongue on the final lateral. With your tongue in this position, try to make an unvoiced stop, being careful to avoid having air pass through the sides of your tongue and 'frictivizing' at your velum. Try to do it with the tip of your tongue between your teeth and you have it pretty much down pat. If you're really brave, you can try pursing your lips as well.

Arabic emphatics are pretty complex sounds, but provided you can get at least one of the 'emphatic' elements right, you should be understood just fine (I once heard a Swahili speaker just use labialization, and it sounded "good enough" to me). Also, despite what a lot of books will tell you, velarization is far more important than pharyngealization. Hell, many neglect to mention the labialization at all.

Now, as for learning to contract your pharynx for /ʕ/ and /ħ/, that's a different story...

Could you (briefly) explain what you mean?

Well, it's true that the "accusative" is used for the direct object of the verb in simple sentances. However, things get more complicated with subordinate clause, where it can often be used for the verbal agent or experiencer (I wouldn't call this ergativity, however). Furthermore, it has various other uses (naked - without prepositions), most of which are adverbial. There are also all kinds of constructions that require the accusative. If you want, I could give some examples (I'm refraining here so as to be brief). Personally, I think "subordinate case" or "adverbial case" would be more appropriate. I brought it up because I've seen people look at constructions and go, "look this noun is in the accusative! It must be the direct object of this verb"! They end up terribly confused and unable to make sense of the sentence as a result.

Well, naturally. I can predict the vowels in most Hebrew words fairly easily, but a lot of that comes from knowing the morphology. As for Arabic, well, as you said the long vowels aren't much of a problem - but I don't have much of an idea about the short vowels yet!

Just thought I'd mention that the Sakhr Dictionary is back up. One feature I like is that it's actually smart enough to peel off inflectional elements and give you nominal/verbal roots. There are times, however, where I feel it isn't thorough enough.

Well, there's one. It's formed with the past tense of "to be" ( = perfect) plus the present participles. It expresses three things:

Compare Arabic:

"was" + active participle = past continuous/progressive
x كان كاتبا /kāna kātiban/ "he was writing"
"was" + imperfect verb = past progressive or iterative
x كان يكتب /kāna yaktubu/ "he was writing, he used to write"

"was" + perfect verb = pluperfect or past conditional
x لو كان له قلم كان كتب /law kāna lahu qalamun, kāna kataba/ "If he had had a pen, he would have written"

As for other tenses, aspects are just implied. Prepositional phrases can be employed if necessary: אני אכתוב את המכתב לא יאוחר מיום שישי 'ani 'echtov 'et hamichtav lo' ye'uchar miyom shishi "I will have written the letter by (lit. "no later than") Friday"

In Arabic, we'd simply use the perfect if the prepositional phrase/adverbs made it clear that the event is taking place in the future. Barring that, we'd use "to be" in the imperfect + main verb in the perfect. Hebrew verbs are definately more "temporal" than Arabic's.

Do you know where the Arabic past tense forms of "be" like kāna are from? That form looks very different from the Hebrew hayah.

Well they most certainly aren't cognates, that's for sure! The whole issue of "be" verbs in Semitic is quite interesting, as they're fairly inconsistant across the family. Most likely, PS simply had no way of expressing 'tense' or 'time' outside of adverbs and adpositional phrases. It was only later, after the family had split, that Semites concerned themself with marking tense on verbs.

As for kāna, the root K-W-N almost certainly dates back to PS, as it appears in East Semitic, where it means something like "firm" or "established": Akkadian /šarru kīnu/ "the true king" (=Sargon). It's not hard to see how such a meaning could change into "being". Actually, I'm fairly certain that Phoenician, which is clearly more closely related to Hebrew than Arabic, also used K-W-N for "to be" instead of Hebrew HYH.

I'm not really certain as to the etymology of hayah, although I do vaguely recall reading that it may have resulted from the verbalizing of /hu'/ and /hi'/. Make of that what you will.

Why is wa- ("and" - I'll install an Arabic keyboard eventually) prefixed?

It's cliticized because it isn't pronounced with a seperate stress.

Does it have any affect on words around it, or do other words affected it?

Hmmm... well, it's always pronounced /wa/. It can, however, cause elision in the following word if it starts with a همزة الوصل /hamzata-l-waṣli/ "glottal stop of liason". An example would be our much discussed استفعل /istafʕala/ verbs: إستعمل قلما /ʔistaʕmala qalaman/ "he used a pen" > واستعمل قلما /wa-staʕmala qalaman/ "and he used a pen".

In Hebrew at least ve- has a number of different forms depending on the structure of whatever it cliticizes to, and in some cases can affect the word itself (eg, ve + Yerushalyim "Jerusalem" > virushalayim).

Colloquially, /wa-/ sometime becomes /ū-/, although I'm not aware of any rules governing the change.

I think you're going to find that Arabic is much more regular than Hebrew, at least with regards to word stems and the like. To look at it from my point of view, Hebrew seems sort of "mushy"; that is a weird 'liquidy' flow of lenitions and changing stress patterns. Arabic seems somehow more 'precise' and 'geometric' to me. Bear in mind that this isn't meant to be a value judgement, just a comment on how fairly closely related languages can sometimes "feel" very different.

Historically-speaking, what does the dotless yaa signify? I realize it's /a:/ in MSA, but why is it used?

I assume you're talking about 'ى' or the ألف مقصورة /'alifun maqṣūratun/ "shortened alif". Apparently, at least some of the words ending in 'ى' were originally pronounced /-ay/, but the sound change occured so late and so inconsistantly that the spelling was kept. Either that, or some dialects pronounced them /-ā/ while others still used /-ay/ and the 'shortened alif' was used as a compromise. I've seen both explanations in various places, although I don't think I've ever seen a discussion in a scholarly context.

In terms of why it's still in use, it actually serves many purposes (off the top of my head).

1. To mark prepositions that change their pronunciation: على /ʕalā/ "on", عليه /ʕalayhi/ "on him".
2. To distinguish between verbs whose last roots are 'y' versus those ending in 'w': مشى /mašā/ "he walked" (M-Š-Y) versus عدا /ʕadā/ "he ran" (ʕ-D-W).
3. To distinguish between indeclinable nouns/adjectives that have seperate emphatic and absolute states مستشفىً /mustašfan/ > المستشفى /al-mustašfā/ "(the) hospital" versus those that don't سينما /sīnamā/ > السينما /as-sīnamā "(the) cinema"

There may be more, but that's all I can think of.

[hwhatting]

I think I should have said "relatively isolated" and not "really isolated". I would argue, however, that all evidence points to the MSA languages existing on the periphery of that continuum. It is probably circular logic, but the fact that the MSA languages didn't go through many of the changes that the "Great West Semitic SprachbundTM" did would tend to point to relative isolation (probably moreso cultural rather than physical) in my mind.

Fair enough - the mere fact that the MSA languages still survive while the areas where OSA languages were most prominently spoken have been arabicised is proof of a certain resilience against Arabic influence.

As for "must have" and "all of their existence", are you sure? For all we know, there were people speaking something related to Basque in South Arabia at some point.

Those were not people. Those were monks.

[Wiseblood]

Fair enough - the mere fact that the MSA languages still survive while the areas where OSA languages were most prominently spoken have been arabicised is proof of a certain resilience against Arabic influence.

MSA and Ethiopian Semitic are very interesting to me, as they both show how much Semitic languages can vary (Chaha in particular completely freaks me out - in a good way ). I often think people don't pay them enough attention.

Those were not people. Those were monks.

Of Course! Clearly, South Arabia was a mere stop off between the Ainu-Dravidian-Basque Urheimat and Europe!

[Cornelius]

Hello gentlemen!

This being my first post, let me start by saying thank you for providing such an thorough introduction and explanation of the workings of Triconsonantal Root-based languages. This thread is fascinating!

I am a bit stumped on one aspect of the root-and-pattern morphology - namely the lack of vowel pattern redundancy. What function prevents future word formation from accidentally making a new word with the same vowel pattern, in effect making a copy of a previously modified base word? (Does that make any sense?) I guess what I'm trying to ask is that with so many different words coming from a root, and so many patterns, you would think there would be 'overlapping' where you get redundancy, but there is none. Is there a specific rule or function preventing this?

Sorry for such a trivial question, but I must be missing something fundamental. Any help would be quite enlightening.

Thanks for any info.

-Cornelius

[Gremlins]

Cornelius: There's quite a strong analogy "shielding" the patterns from sound change; a conditioned change would either be treated as allophony or revert back by analogy. Some conditioned changes like Hebrew lenition get through iirc because they don't merge many roots. Intervocalic voicing of voiceless stops, for example, could cause serious havoc.

[Mecislau]

This got me to thinking about how we tend to naturalize foreign words in Arabic. As an example, the Classical Arabic word for "orange" (the fruit) is برتقال /burtaqālu/ (clearly a greek loan). This is decidedly "un-semitic". Over time the word changed to become more in line with other noun patterns:

burtaqālu
*burtaqāl (loss of case ending)
*burta'āl ( /q/ > /ʔ/)
*burt'āl (unstressed vowel is dropped)
*burṭāl (glottal stop is dropped to avoid consonant cluster, /t/ becomes emphatic due to glottal influence)
*burṭān (/l/ > /n/)

I've definitely noticed this before. While researching a lot of Russian etymologies, I found a number of couplets of Greek words borrowed more or less directly into Russian alongside the same Greek word that passed through Arabic first. These words, though, seem to look completely different. When looking back through the history of the Arabic loan, it turns out that every other language borrowed the word with few modifications, except for Arabic, which completely butchered it

For example, compare Russian адамант adamant (from Greek adamantos) with алмаз almaz (from Arabic الْماس 'almaas).

The final form, burṭān looks like any other word of the fuʕlān type (bustān "garden", burdān "Cold").

Although Hebrew doesn't have have any foreign loans that have undergone such huge reductions, it does have a number of more recent coinages that are made with native morphology and vocabulary, but appear reminiscent of foreign words - perhaps one step more native than borrowing.

Two examples off the top of my head:

אווירון 'aviron "airplane", formed from אוויר 'avir "air" + the "tool/machine" suffix -on. The result is a native word that looks quite like French avion or Spanish avión. (This word is a bit dated, however - the completely native word מטוס matos, from the root T-W-S "fly", is more common nowadays).

מקסים maqsim "charming, enchanting". I'm not sure if this is a foreign-influenced word or not, but it feels like it could be to me (such as the Latinate maxim-). Derived from the root Q-S-M "charm, enchant", using a normal hif`il present participle.

Interesting, my understanding is that Maltese does similar things. I think it'd be rather interesting to do a study comparing European influences on both Modern Hebrew and Maltese.

My understanding is that Maltese has undergone a great deal more Europeanization than Hebrew has, though, and now has a huge set of words that don't follow any sort of consonantal morphology.

Also, quick comment: I was looking on the Hebrew Academy website to see if they had come up with any natively-formed word for "globalization", and I came across this completely unrelated jewel of a word: וווו uvavo "and his hook", formed with four vavs. Of course, this is the kind of word that will confuse any native speaker when they first see it...

Well, it's true that the "accusative" is used for the direct object of the verb in simple sentances. However, things get more complicated with subordinate clause, where it can often be used for the verbal agent or experiencer (I wouldn't call this ergativity, however). Furthermore, it has various other uses (naked - without prepositions), most of which are adverbial. There are also all kinds of constructions that require the accusative. If you want, I could give some examples (I'm refraining here so as to be brief). Personally, I think "subordinate case" or "adverbial case" would be more appropriate. I brought it up because I've seen people look at constructions and go, "look this noun is in the accusative! It must be the direct object of this verb"! They end up terribly confused and unable to make sense of the sentence as a result.

I'd love to see some examples. This sounds very interesting!

(With Hebrew, of course, I have very little idea of how cases in Semitic languages work. Hebrew's completely lost cases, with only a few remnants remaining)

A question for you: What's the most commonly use Arabic word for "night"? The Sakhr dictionary shows both layl and laylah. There is a similar situation in Hebrew, where both לילה laylah (more common) and ליל lel exist (and also ליל layil, actually, although that's more poetic). Do you have any idea where this came from?

I've heard it explained that לילה (which is a masculine noun, and declines as though it were layil) was an example of an accusative case ending that became merged with the noun for whatever reason. Laylah is indeed the form an accusative layil would take in Hebrew (compare bayit "house" and its accusative (ha)baytah "homewards" - the "accusative" in modern Hebrew is used as a directional), but if Arabic has this variation too, then this merger must have occurred very early on, or there is another explanation.

I'm not really certain as to the etymology of hayah, although I do vaguely recall reading that it may have resulted from the verbalizing of /hu'/ and /hi'/. Make of that what you will.

It's a fun root for sure: H-Y-H. Three particularly common gzarot (paradigms for roots containing certain problematic/defective consonants) are root-initial H, root-medial Y, and root-final H

Actually, oddly enough, this particular verb is surprisingly regular given the above. Of course, no one actually pronounces it the way it's "supposed" to be pronounced - in the future tense the "proper" forms ('ehyeh, tihyeh, tihyi, etc) contain consonant clusters that are actually illegal in Hebrew, so people either drop the /h/ (saying 'eye, tiye, tiyi, ...) or add in an /i/: ('ehiye, tihiye, tihiyi, ...).

I could definitely see this being related to pronouns like hu' and hi'. That might also help explain why the verb isn't nearly as irregular as would be expected.

I think you're going to find that Arabic is much more regular than Hebrew, at least with regards to word stems and the like. To look at it from my point of view, Hebrew seems sort of "mushy"; that is a weird 'liquidy' flow of lenitions and changing stress patterns. Arabic seems somehow more 'precise' and 'geometric' to me. Bear in mind that this isn't meant to be a value judgement, just a comment on how fairly closely related languages can sometimes "feel" very different.

Hmm, interesting. In some respects I see what you mean, but in others I see the opposite. For example, (Modern Standard) Arabic's words all seem to run together to me, as I've already seen many instances where one word affects the pronunciation of others nearby, such as the hamzata-l-waṣli (which is triggered by many different words), whereas in Hebrew the words feel much more distinct and separate from one another.

I'm not entirely sure what you mean by changing stress patterns, though. Hebrew's very consistent, always having word-final stress except in a few predictable word forms. Arabic's stress system seems a lot more confusing to me! Then again, I've spoken Russian and English my whole life, both of which are known for painful stress systems...

In respect to lenition and overall regularity, though, I definitely see what you mean.


Does Arabic make heavy use of reduplicated roots? Hebrew has a large set of 4-consonantal roots that consist of a reduplicated biconsonantal root, often onomatopoeia: צלצל tziltzel "ring", קשקש qishqesh, בלבל bilbel "confuse", פטפט pitpet "chatter", זלזל zilzel "scorn", etc. You also see this in some nouns, such as פלפל pilpel "pepper". Hebrew also uses reduplication of the final two consonants of a root to form diminutives (חתול chatul "cat" > חתלתול chataltul "kitten") or to blur color terms (אפור 'afor "gray" > אפרפר 'afarpar "grayish").

I am a bit stumped on one aspect of the root-and-pattern morphology - namely the lack of vowel pattern redundancy. What function prevents future word formation from accidentally making a new word with the same vowel pattern, in effect making a copy of a previously modified base word? (Does that make any sense?) I guess what I'm trying to ask is that with so many different words coming from a root, and so many patterns, you would think there would be 'overlapping' where you get redundancy, but there is none. Is there a specific rule or function preventing this?

I'm sorry, I don't quite understand what you mean. Could you maybe try to provide an example?


Identical vowel patterns do exist, but they're generally shared across multiple parts of speech. CaCaC, for example, is the third person singular past tense of verbs in the pa`al group, the masculine singular form of a number of adjectives, and the singular form of some nouns. None of these three patterns are related. There's rarely confusion, however, because in a sentence it's generally not hard to distinguish a noun from an adjective from a verb.

[Cornelius]

Sorry if my question was a bit vague.

Lets say I need to make a new word for "notepad". What prevents the word from taking a previously made word and making an inadvertant clone? For example, "katav", where the verb form 'katav' "He wrote" (3Sg Masc Past) is redundantly recreated in 'katav' "journalist, reporter". I understand that context can take care of most of the problems, but you would think with such a high number of forms and a limited number of vowels the chance of redundancy could cause a lot of problems.

Now I'm fairly sure the above example is an exception, and not the norm, so what prevents the redundancy?

*EDIT*
I have done some more in-depth reading of Hebrew grammar, and I think I understand a bit better. Is pattern redundancy eliminated (almost) in noun formation by the use of applying various "meters" to the same roots in the same way that Hebrew verbs are conjugated by applying various prefixes, suffixes and internal vowel combinations?

Thanks for your patience, my knowledge of Semitic languages is nearly nil (and I received poor marks in English as well, and thats my native language!).
*EDIT*

-Cornelius

[Mecislau]

Sorry if my question was a bit vague.

Lets say I need to make a new word for "notepad". What prevents the word from taking a previously made word and making an inadvertant clone? For example, "katav", where the verb form 'katav' "He wrote" (3Sg Masc Past) is redundantly recreated in 'katav' "journalist, reporter". I understand that context can take care of most of the problems, but you would think with such a high number of forms and a limited number of vowels the chance of redundancy could cause a lot of problems.

Now I'm fairly sure the above example is an exception, and not the norm, so what prevents the redundancy?

*EDIT*
I have done some more in-depth reading of Hebrew grammar, and I think I understand a bit better. Is pattern redundancy eliminated (almost) in noun formation by the use of applying various "meters" to the same roots in the same way that Hebrew verbs are conjugated by applying various prefixes, suffixes and internal vowel combinations?

Thanks for your patience, my knowledge of Semitic languages is nearly nil (and I received poor marks in English as well, and thats my native language!).
*EDIT*

-Cornelius

Basically, yes, that's it. The vowel patterns used to derive nouns (in English often called meters, in Hebrew mishqalim "weights") each have their own meaning, just as derivational affixes in English such as -er, -tion, -th, etc, have. The pattern CaCaC is much like -er in English, forming nouns indicating profession: katav "reporter" (< KTB "write"), tabach "chef" ( < TBCh "cook"), tayas "pilot" ( < TVS "fly" - Hebrew has a lot of Y/V alternation, since the two phonemes /j/ and /w/ were confused early on), etc.

Now this doesn't mean all CaCaC nouns are professions, but a large number of them certainly are.

"Notebook", however, probably would not use this pattern. It's a thing, not a person. The word in Hebrew is machberet, derived from the root ChBR (generally meaning "join" or "bind", but also meaning "compose [writing]") using the mishqal miCCeCet - the vowels vary slightly because of the presence of the sound /X/, which tends to draw vowels around it toward /a/. This pattern is shared by such nouns as miqlachat "shower (stall)" (< QLCh "shower"), mizchelet "sled, sledge" ( < ZChL "crawl"), etc.




Wiseblood: Three more foreign affixes in Hebrew are -ונר -oner and -צ'יק -tshiq, and -ניק -niq, again all derived from Russian, but are far less common than -atzya. These three all form agentives.

[Cornelius]

@ Maknas: Thank you very much! That clears things up nicely. Tricon root word formation seems a bit 'clunky' at first to an English speaker, but is actually very elegant once the nuts and bolts are revealed...truly fascinating!

Cornelius: There's quite a strong analogy "shielding" the patterns from sound change; a conditioned change would either be treated as allophony or revert back by analogy. Some conditioned changes like Hebrew lenition get through iirc because they don't merge many roots. Intervocalic voicing of voiceless stops, for example, could cause serious havoc.

I must be in the deep end of the pool... =)

-Cornelius

[Wiseblood]

Cornelius: There's quite a strong analogy "shielding" the patterns from sound change; a conditioned change would either be treated as allophony or revert back by analogy. Some conditioned changes like Hebrew lenition get through iirc because they don't merge many roots. Intervocalic voicing of voiceless stops, for example, could cause serious havoc.

It's funny you should use voicing as an example of "havoc creation", as there are a series of Gurage languages that have replaced consonant gemination with devoicing! Actually, Gurage (especially Chaha) is kind of famous for doing all kinds of crazy mutations with consonant roots.

I've definitely noticed this before. While researching a lot of Russian etymologies, I found a number of couplets of Greek words borrowed more or less directly into Russian alongside the same Greek word that passed through Arabic first. These words, though, seem to look completely different. When looking back through the history of the Arabic loan, it turns out that every other language borrowed the word with few modifications, except for Arabic, which completely butchered it.

Heh, it's funny, but some of those Latin/Greek borrowings were so heavily naturalized that Hebrew and Aramaic words look more foreign.

Although Hebrew doesn't have have any foreign loans that have undergone such huge reductions, it does have a number of more recent coinages that are made with native morphology and vocabulary, but appear reminiscent of foreign words - perhaps one step more native than borrowing.

Compare colloquial Arabic موت سريع /mōt sariʕ/ "motorcycle" (lit. "a quick death").

My understanding is that Maltese has undergone a great deal more Europeanization than Hebrew has, though, and now has a huge set of words that don't follow any sort of consonantal morphology.

Maltese is almost certainly more europeanized than MH. I was just making the point that it would be interesting to see how two seperate Semitic languages have coped with "external linguistic pressure".

What I find hilarious about Maltese is that I can understand it better than, say, Moroccon Arabic. It probably helps that I speak French. Then again, there are those that'd argue that speaking French should help with understanding Moroccon Arabic too.

Also, quick comment: I was looking on the Hebrew Academy website to see if they had come up with any natively-formed word for "globalization", and I came across this completely unrelated jewel of a word: וווו uvavo "and his hook", formed with four vavs. Of course, this is the kind of word that will confuse any native speaker when they first see it...

I've got a better (or at least more esoteric) one: يييّي /yuyayyī/ "he writes a beautiful 'yā'".

I'd love to see some examples. This sounds very interesting!

Off the top of my head:

Direct object of verb:
ضرب حسنٌ سليماً
ḍaraba ħasanun salīman
hit-he Hassan-nom Saleem-acc
"Hassan hit Saleem"

As the subject of a subordinate clause:
قالوا إنّ سليماً قتل حسناً
qālū 'inna salīman qatala ħasanan
say-they that Saleem-acc killed-he Hassan-acc
"They say that Saleem killed Hassan"

This is even true with passive verbs:
قالوا إنّ حسناً قُتل
qālū 'inna ħasanan qutila
say-they that Hassan-acc (was killed)-he
"They say that Hassan was killed"

Predicate in existential and inchaotive verbs:
كان سليمٌ مجنوناً
kāna salīmun majnūnan
was-he Saleem-nom crazy-acc
"Saleem was crazy"

أصبح جبَّاراً.
'aṣbaħa jabbāran
became-he tyrant-acc
He became a tyrant.

As the subject in emphatic sentences (notice how subject and predicate reverse cases):
إنّ سليماً مجنونٌ
'inna salīman majnūnun
verily Saleem-acc crazy-nom
"Saleem is very crazy"


Several Adverbial uses:

As المفعول المطلق /al-mafʕūlu-l-muṭlaqu/ "the absolute object", discussed earlier:
ضربه ضرباً شديداً
ḍarabahu ḍarban šadīdan
hit-he-him blow-acc intense-acc
"he struck him a strong blow" = "he hit him hard"

Expressing purpose:
حكيتُ سكوتاً له
ħakaytu sukūtan lahu
spoke-I (shutting-up)-acc for-him
"I spoke to shut him up"

To express time (remember اسم المرّة /ismu-l-marrati/ "the noun of time" from a while back?):

جلستُ دقيقتين
jalastu daqīqatayn
sat-I minute-fem-(dual oblique)
"I sat for two minutes"

نمتُ صباحاً كاملاً
nimtu ṣabāħan kāmilan
slept-I morning-acc complete-acc
"I slept all morning"

To express what we call حال /ħālun/ "state":
أتى سائقاً
'atā sā'iqan
came-he driving-acc
"he came driving"

تحادثوا ماشين
taħādaθū māšīna
(self-talked)-they walking-(plural oblique).
"They talked while walking"

رأيت الباب مفتوحاً
ra'aytu-l-bāba maftūħan
saw-I the door-acc opened-acc
"I saw the door open"

To make matters more confusing, a ħāl can take its own object:
أتى سائقاً سيَّارةً
'atā sā'iqan sayyāratan
came-he driving-acc car-fem-acc
"he came driving a car"

For تمييز /tamyīzun/ which means something like "specifying" or "differentiating":
هي أكبر مدينة سكَّاناً
hiya 'akbaru madīnatin sukkānan
she bigger city-fem-gen inhabitants-acc
"it's the biggest city with respect to population"

Replacing prepositions with verbs of motion (see below):
ذهب الى البيت/ ذهب البيت
ðahaba 'ila-l-bayti / ðahaba-l-bayta
went-he to the house-gen / went-he the house-acc
"he went (to) the house"

خرجت من البيت / خرجت البيت
xarajat mina-l-bayti / xarajati-l-bayta
left-she from the house-gen / left the house-acc
"she exited (from) the house"

As the indirect object with verbs of motion + /bi/ "with":

أتى سليمٌ حسناً بكتابين
'atā salīmun ħasanan bikitābayni
came-he Saleem-nom Hassan-acc with book-(dual oblique)
"Saleem brought two books to Hassan"

Partatively, after 11-99 as well as with كم /kam/ "how many?":
كم كتاباً قرأت؟
kam kitāban qara'ta
(how many) book-acc read-you
"how many books did you read?"
قرأت سبعة عشر كتاباً.
qara'tu sabʕata ʕašara kitāban
read-I seven-(acc cons) ten-(acc cons) book-acc
"I read seventeen books"

As a bit of a side note, a great many "prepositions" in Arabic are really just nouns in the accusative construct:
x خلِف /xalafa/ "he followed" > خَلْف /xalfa/ "behind"
x قبِل /qabila/ "he met" > قَبْل /qabla/ "before"
x فاق /fāqa/ "he bettered" > فوق /fawqa/ "above"

There are almost certainly more that I'm missing, but I think I may have proved my point.

A question for you: What's the most commonly use Arabic word for "night"? The Sakhr dictionary shows both layl and laylah. There is a similar situation in Hebrew, where both לילה laylah (more common) and ליל lel exist (and also ליל layil, actually, although that's more poetic). Do you have any idea where this came from?

Well, I don't know for sure about Hebrew, but in Arabic, ليل /laylun/ refers to night in general while ليلة /laylatun/ refers to a "specfic" night". Compare بقر /baqara/ "cow(s)" (in general) > بقرة /baqaratun "a cow" > بقرات baqarātun "cows". It's possible that the Hebrew forms are holdovers of something similar.

I've heard it explained that לילה (which is a masculine noun, and declines as though it were layil) was an example of an accusative case ending that became merged with the noun for whatever reason. Laylah is indeed the form an accusative layil would take in Hebrew (compare bayit "house" and its accusative (ha)baytah "homewards" - the "accusative" in modern Hebrew is used as a directional), but if Arabic has this variation too, then this merger must have occurred very early on, or there is another explanation.

See above about accusatives and verbs of Motion in Arabic. What's really happening is that the "accusative" isn't acting as an "accusative" but as an "adverbial" case, so in effect you're really saying "I'm going homewards", and not "*I'm going house". Interestingly enough, Akkadian had a seperate "adverbial" case, ending in -iši. Assuming that -h at the end of the examples above was actually pronounced, the Hebrew form could just as easily be descended from that.

Hmm, interesting. In some respects I see what you mean, but in others I see the opposite. For example, (Modern Standard) Arabic's words all seem to run together to me, as I've already seen many instances where one word affects the pronunciation of others nearby, such as the hamzata-l-waṣli (which is triggered by many different words), whereas in Hebrew the words feel much more distinct and separate from one another.

When I look at a Hebrew/Aramaic word, it's like someone decided to randomly lengthen and shorten vowels all over the place, all the while finding new and interesting ways of forming consonant clusters. I actually think this may have something to do with the way Semitic speakers see word "contours"; that is Hebrew word "contours" seem somehow "out of shape" to me. I don't know if this makes any sense.

I'm not entirely sure what you mean by changing stress patterns, though. Hebrew's very consistent, always having word-final stress except in a few predictable word forms.

In terms of stress patterns, I meant historical stress patterns. Again, I often find myself asking, "why is this vowel long?" and "why was that syllable deleted?" But yeah, MH does have pretty regular stress as I understand it.

Arabic's stress system seems a lot more confusing to me! Smile Then again, I've spoken Russian and English my whole life, both of which are known for painful stress systems...

Heh, I have no idea about Russian, but Arabic's stress system is cake compared to English's (mostly because it's regular). Stress falls on the penultimate if it's long. Otherwise, it falls on the antepenultimate. A long syllable is either CV: or CVC. That's it!

Does Arabic make heavy use of reduplicated roots? Hebrew has a large set of 4-consonantal roots that consist of a reduplicated biconsonantal root, often onomatopoeia: צלצל tziltzel "ring", קשקש qishqesh, בלבל bilbel "confuse", פטפט pitpet "chatter", זלזל zilzel "scorn", etc. You also see this in some nouns, such as פלפל pilpel "pepper". Hebrew also uses reduplication of the final two consonants of a root to form diminutives (חתול chatul "cat" > חתלתול chataltul "kitten") or to blur color terms (אפור 'afor "gray" > אפרפר 'afarpar "grayish").

Yes, very much so: زلزل /zalzala/ "tremble", دقدق /daqdaqa/ "knock", قهقه /qahqaha/ "laugh loudly", بلبل /balbala/ "confuse", وسوس /waswasa/ "whisper/ ...and my personal favorite, بأبأ /ba'ba'a/ "call someone father". أبأبئني /'uba'bi'nī/ "call me daddy!"

Basically, yes, that's it. The vowel patterns used to derive nouns (in English often called meters, in Hebrew mishqalim "weights") each have their own meaning, just as derivational affixes in English such as -er, -tion, -th, etc, have. The pattern CaCaC is much like -er in English, forming nouns indicating profession: katav "reporter" (< KTB "write"), tabach "chef" ( < TBCh "cook"), tayas "pilot" ( < TVS "fly" - Hebrew has a lot of Y/V alternation, since the two phonemes /j/ and /w/ were confused early on), etc.

Can I guess that from the /b/ in the middle of the word that tabach was once *tabbach? If that's the case, than that would be very close to the Arabic "professional" form, طبّاخ /ṭabbāxun/ "chef".

Continued Below

[Wiseblood]

I was wondering when I'd reach the post-length-limit. I guess I just found out.

"Notebook", however, probably would not use this pattern. It's a thing, not a person. The word in Hebrew is machberet, derived from the root ChBR ...

Heh, one of the more common words for "notebook" in Arabic دفتر /daftarun/, definitely foreign (Persian, I think). Another one is تأليف /ta'līfun/, from ألّف /'allafa/ "he penned, scribed". This bring up another point about vocab building in Semitic languages: just because you have one root meaning ie. writing (ktb), doesn't mean there aren't others that overlap in places. There are all kinds of roots that "writing" or "literature" or "drawing" all of which could be applied to this situation (ktb,ʔlf,sṭr,ʔdb,šʕr).

Also, I imagine Hebrew, like Arabic, has more than one form for different word "categories": /miftāħun/ "key" > /fattāħun/ "can opener". There probably half a dozen forms in Arabic that mean "tool".

[hwhatting]

On the uses of the Accusative:
Actually, the only cases that really look exotic from an Indo-European point of view are the uses for the predicate, the one for the emphatic subject, and the use after certain numerals. Except for those, you'd find similar uses of the accusative to what you listed in many IE languages, e.g., Latin:

As the subject of a subordinate clause:
قالوا إنّ سليماً قتل حسناً
qālū 'inna salīman qatala ħasanan
say-they that Saleem-acc killed-he Hassan-acc
"They say that Saleem killed Hassan"

Latin could use an AcI (accusativus cum infinitivo) here:
Marcum Lucium interfecisse dicunt.
Marcus-acc Lucius-acc kill-inf.perf. say-3pl.

This is even true with passive verbs:
قالوا إنّ حسناً قُتل
qālū 'inna ħasanan qutila
say-they that Hassan-acc (was killed)-he
"They say that Hassan was killed"

Lucium interfectum esse dicunt.
Lucius-acc kill-past.passive.participle-acc.sg.m. be-infinitive say-3pl.

Granted, Arabic tops this by having finite verb forms, but I'd say the principle is similar - you just need to analyse the acc. for the subject of the dependent clause as being dependent of the "say" in the main clause.

For the adverbial uses, cf. German for the accusative of time:

تُ دقيقتين
jalastu daqīqatayn
sat-I minute-fem-(dual oblique)
"I sat for two minutes"

Ich saß zwei Minuten.
I sat two minutes

نمتُ صباحاً كاملاً
nimtu ṣabāħan kāmilan
slept-I morning-acc complete-acc
"I slept all morning"

Ich habe den ganzen Morgen geschlafen.
I have the-acc.sg.m. whole-acc.g.m. morning(-acc.sg.) slept.


So, these things may come as a surprise for speakers of English or a Romance language, but they won't be as surprising for someone who knows the Classical IE languages or a modern IE language with a full case system.

[Mecislau]

What I find hilarious about Maltese is that I can understand it better than, say, Moroccon Arabic. It probably helps that I speak French. Then again, there are those that'd argue that speaking French should help with understanding Moroccon Arabic too.

Heh, yes, I've heard that Moroccan Arabic is very divergent compared to the other varieties of Arabic. The only expression I remember from when I was there a few years back that was unique was the use of zhuzh for "two" instead of the standard 'ithnāni (which I'm guessing is cognate to Hebrew זוג zug "pair").

I've got a better (or at least more esoteric) one: يييّي /yuyayyī/ "he writes a beautiful 'yā'".

Is that a real word?!

Off the top of my head:

Like hwhatting said above, a lot of these aren't that strange, except when you use the accusative for the subject. Your emphatic example is just weird, though. In a good way.

Replacing prepositions with verbs of motion (see below):
ذهب الى البيت/ ذهب البيت
ðahaba 'ila-l-bayti / ðahaba-l-bayta
went-he to the house-gen / went-he the house-acc
"he went (to) the house"

Yeah, this is basically the only place where the accusative ending survives in modern Hebrew, and even then only in some very limited circumstances:

הוא הלך הביתה
hu' halach habáytah
he went-he the-house-acc
"He went home"

In Biblical Hebrew the accusative ending (in this same directional function) was much more productive.

However, it can only be used in expressing motion toward, not away from as in your "she exited from the house" example:

היא יצאה מהבית
hi' yatz'ah mehabáyit
she left-she from-the-house
"She left the house"

The nominative ending -u was completely dead by Biblical Hebrew, even, with only a few traces left behind. The only one that comes to mind right off hand is the Biblical name Bethuel (BH בתואל bəthu'el), meaning "house of God", which in normative Biblical Hebrew would be בת-אל beth-'el.

However, given that the nominative -u seemed to cause the /e/ in the construct "beth" (which already is reduced from /ayi/ in the absolute "bayith") to reduce further to schwa, I'm quite glad this didn't survive!

See above about accusatives and verbs of Motion in Arabic. What's really happening is that the "accusative" isn't acting as an "accusative" but as an "adverbial" case, so in effect you're really saying "I'm going homewards", and not "*I'm going house". Interestingly enough, Akkadian had a seperate "adverbial" case, ending in -iši. Assuming that -h at the end of the examples above was actually pronounced, the Hebrew form could just as easily be descended from that.

Nope, that /h/ was never pronounced. It's just an orthographic device. I've just gotten into the habit of transcribing it as well.

Hebrew only has a true word-final /h/ in a very small set of words/forms - the 3sg feminine singulative possessive/prepositional ending -ah (לה lah "to her", ספרה sifrah "her book") and in a small set of other words (גבוה gavóah "high, tall"). In modern Hebrew these /h/s are rarely pronounced, though.

Hmm, interesting. In some respects I see what you mean, but in others I see the opposite. For example, (Modern Standard) Arabic's words all seem to run together to me, as I've already seen many instances where one word affects the pronunciation of others nearby, such as the hamzata-l-waṣli (which is triggered by many different words), whereas in Hebrew the words feel much more distinct and separate from one another.

When I look at a Hebrew/Aramaic word, it's like someone decided to randomly lengthen and shorten vowels all over the place, all the while finding new and interesting ways of forming consonant clusters. I actually think this may have something to do with the way Semitic speakers see word "contours"; that is Hebrew word "contours" seem somehow "out of shape" to me. I don't know if this makes any sense.

Not really ( ), but I do sorta get what you mean. Arabic, Amharic, and many other Semitic languages underwent very few phonetic changes from Proto-Semitic, and most of what they did change was unconditional (ie, all X becomes Y). Hebrew had a load of conditional vowel changes that sent the vowels all over the place, due to stress, environment, other nearby vowels...

Heh, I have no idea about Russian, but Arabic's stress system is cake compared to English's (mostly because it's regular). Stress falls on the penultimate if it's long. Otherwise, it falls on the antepenultimate. A long syllable is either CV: or CVC. That's it!

...

You just explained the Arabic stress system in a single sentence, while my book took an entire page for it, and you did it clearer.

Thanks

Can I guess that from the /b/ in the middle of the word that tabach was once *tabbach? If that's the case, than that would be very close to the Arabic "professional" form, طبّاخ /ṭabbāxun/ "chef".

Yes. Almost always when you see a non-initial /b p k/ in modern Hebrew, that comes from a former geminate /bb pp kk/, since their non-geminate forms all lenited.

Biblical Hebrew still had geminates, while modern Hebrew does not. However, if you keep this in mind a number of other similarities between Hebrew and Arabic show up. For example, Biblical Hebrew did have the geminate middle consonant in the intensive paradigm: dibber "he said" (modern diber). The first consonant of a word also underwent gemination when the definite articles to added: BH bayith "house", habbayith "the house", the same sort of gemination you see with the Arabic article and the sun letters.

I was wondering when I'd reach the post-length-limit. I guess I just found out.

I never realized there was one!

Also, I imagine Hebrew, like Arabic, has more than one form for different word "categories": /miftāħun/ "key" > /fattāħun/ "can opener". There probably half a dozen forms in Arabic that mean "tool".

Well, naturally. The language would be quite limited without that.

מפתח maftéach "key"
פותחן potchan "can opener"

Those forms look quite similar to the Arabic...


Do you happen to know of any good resources on Proto-Semitic grammar? Everything I can find is very scattered about. I'm particularly interested in PS verbal formation, since it seems like the different Semitic languages have taken it various different directions.


The construct in different languages (and PS) also interests me a bit now, after looking into how Arabic forms it. When the head (construct) noun is singular, Hebrew and Arabic are quite similar - masculine nouns are endingless and indefinite, and feminine nouns regain the -t that they had lost elsewhere. When the head noun is plural, though, the two languages go completely different directions:

In Hebrew, plural feminine construct nouns are almost identical to their absolute forms, but generally have some reduced vowels (eg, 'aratzot "lands (abs)" > 'artzot "lands of (cons)"). Masculine nouns, however, take a distinct ending -ey: sfarim "books" > sifrey "books of".

Many singular construct nouns undergo changes as well, such as báyit "house" > beyt "house of".

Arabic, on the other hand, appears to drop the /n/ ending in regular externally-derived plurals: mu`allimûna "teachers" > mu`allimû "teachers of".

Any idea where these might be coming from?



Hmm, okay, even as I wrote that I think I may have figured it out. The Hebrew process is the same as the Arabic, but is a little obscured by sound changes. Using a more regular noun as an example, you have words like po`alîm "workers" that at originally just lost the /m/ in construct, becoming po`alî "workers of". A regular sound change then shifted that final long /i:/ to /e:/, giving Biblical and then modern po`aley.

Did Proto-Semitic form construct plurals by just lengthening the case ending vowel of the singular, then?


I'm really beginning to like this comparative Semitic stuff!

[Wiseblood]

On the uses of the Accusative:
Actually, the only cases that really look exotic from an Indo-European point of view are the uses for the predicate, the one for the emphatic subject, and the use after certain numerals. Except for those, you'd find similar uses of the accusative to what you listed in many IE languages, e.g., Latin:

Well, this wasn't meant to be a list of "strange Arabic accusative constructionsTM", so much as a list that people could poke at. I showed these to my wife, who happens to know German and Latin as well. Granted, both are pretty rusty for her, and she's not a native German speaker (for that matter, she's not a native Latin speaker either. ). However, she did say that some of these hurt her brain.

Marcum Lucium interfecisse dicunt.
Lucium interfectum esse dicunt.

Interesting (bear in mind I haven't studied Latin in almost ten years now). Tell me, is this construction valid for any subordinate clause, or does it only apply to, say, "indirect reporting" verbs? What about existential/inchaotive constructions? For example, in Arabic, you can say:

من الممكن أنّ سليماً قتل حسناً
mina-l-mumkin.i 'anna salīm.an qatal.a ħasan.an
from-the-possible.gen that Saleem.acc killed.3ms Hassan.acc
"It's possible that Saleem killed Hassan"

لا شكّ من أنّ سليماً مجنونٌ
lā šakk.a min 'anna salīm.an majnūn.un
no doubt.acc from that Saleem.acc crazy.nom
"There is no doubt that Saleem is crazy" (Obviously because he killed poor Hassan )

I'm actually quite surprised neither you nor Maknas found أتى سليمٌ حسناً بكتابين "Saleem brought two books to Hassan" strange as I've had people tell me that "came X Y with Z" confuse them. However, my wife thought it was merely a double accusative verb, as the dual only has nominative and oblique forms. She suggested I give a less ambiguous example:

أتى بكتابٍ سليمٌ حسناً
'atā.0 bi.kitāb.in salīm.un ħasan.an
came.1ms with.book.gen Saleem.nom Hassan.acc
Saleem brought a book to Hassan.

Ich saß zwei Minuten.
Ich habe den ganzen Morgen geschlafen.

As I said, it's been a while since I studied Latin, but even I can remember constructions like noctem totem dormivi. No, I couldn't remember the Latin word for Morning, sue me.

Heh, yes, I've heard that Moroccan Arabic is very divergent compared to the other varieties of Arabic. The only expression I remember from when I was there a few years back that was unique was the use of zhuzh for "two" instead of the standard 'ithnāni (which I'm guessing is cognate to Hebrew זוג zug "pair").

Well, North African Arabic in general seems pretty strange to me (except for Libyan, where I can almost make out a full sentence every now and then ). My understanding is that Maghrebi Arabic went through a pretty severe sound shift in terms of vowels - something like: /a, i, u/ > /ə/ then /ā, ī, ū/ > /a, i, u/ and finally, /ay, aw/ > /ī, ū/. They also seem affricate their dental stops all over the place. I'm probably simplifying a fair bit, but I think you get the idea.

So far as I know, žūž does come from CA زوج /zawjun/ "double" or "spouse" (the verb زوّج /zawwaja/ means "he married")

Is that a real word?!

Believe me when I tell you, I couldn't make that one up if I was on acid. It is a real word, just a really esoteric one.

The nominative ending -u was completely dead by Biblical Hebrew, even, with only a few traces left behind. The only one that comes to mind right off hand is the Biblical name Bethuel (BH בתואל bəthu'el), meaning "house of God", which in normative Biblical Hebrew would be בת-אל beth-'el.

Heh, the retention of /-u/ in the construct is fairly common in colloquial Arabic as well. For the record, "house of God" in Arabic would be, بيت الله /baytu-llāhi/

Not really ( Razz ), but I do sorta get what you mean. Arabic, Amharic, and many other Semitic languages underwent very few phonetic changes from Proto-Semitic, and most of what they did change was unconditional (ie, all X becomes Y). Hebrew had a load of conditional vowel changes that sent the vowels all over the place, due to stress, environment, other nearby vowels...

I'd make the arguement that Amharic has undergone some pretty severe sound changes from PS: *wal(a)dum > ləǧ "boy". I will agree, however that Amharic word "shapes" seem a little more familiar to me: Ar: kāhinun, Amh: kahən vs He: kōhēn - Ar: kawākibu, Amh: käwakəbt vs. He: *kōkābīm (not too sure on the vowels for this). Transliteration differences aside, the Arabic and Amharic words are pronounced almost identically.

Then again, Amharic's loss of gutterals can reverse this somewhat: He: ħādāš seems a lot more like Ar: ħadīθun to me than does Amh: addis.

You just explained the Arabic stress system in a single sentence, while my book took an entire page for it, and you did it clearer.

Only a page? I can't count the number of times I've read some boroque three-page explanation of Arabic stress involving counting the syllables backwards or somesuch blah blah. Granted, there are different schools of thought as to how to stress CA, but the method I gave you is by far the most common, and is easier to reconcile with Modern Arabic dialects.

Biblical Hebrew still had geminates, while modern Hebrew does not. However, if you keep this in mind a number of other similarities between Hebrew and Arabic show up. For example, Biblical Hebrew did have the geminate middle consonant in the intensive paradigm: dibber "he said" (modern diber). The first consonant of a word also underwent gemination when the definite articles to added: BH bayith "house", habbayith "the house", the same sort of gemination you see with the Arabic article and the sun letters.

North Arabian languages had haC:-/han- for the definite article as well. Given the prominence of nasal-assimilation in Hebrew, some people have suggested a Proto-Caano-Arabic definite article, *han-, with the changes /h/ > /ʔ/ and /n/ > /l/ occuring in Arabic. Actually, Semitic "intensive" forms like *paʕʕala are supossedly come from Proto-Afro-Asiatic *panʕala. You can see this today in Beja (not Semitic, but Afro-Asiatic) verbs: *aktib "I write" > *akantib "I am writing" (bear in mind, I'm not certain if the vowels are correct). Apparently, Aramaic used to "randomly" alternate between /C:/ and /nC/ in such verb forms as well.

Those forms look quite similar to the Arabic...

Yeah, مفتاح /miftāħun/ and מפתח maftéach are clear cognate forms. I actually screwed up the second one, though; it should be فتّاحة /fattāħatun/ and not فتّاح /fattāħun/.

Do you happen to know of any good resources on Proto-Semitic grammar? Everything I can find is very scattered about. I'm particularly interested in PS verbal formation, since it seems like the different Semitic languages have taken it various different directions.

You know, I don't think I've ever seen a really good PS grammar (most of what I know I've collected in bits and pieces from various sources). There's Lipinski's Semitic Languages: Outline of a Comparative Grammar, which has a fair bit of information scattered about. Now it's a comparative outline, so it's not really a thorough reconstruction or anything. Also, I found myself questioning his theories a fair bit. He also seems to have an unhealthy fetish for Berber. Then again, people like Ehret take it to far in the other direction: "Yeah, well, I think I'm just going to ignore Berber in my reconstruction of Proto-Afro-Asiatic, even though next to Semitic and Egyptian, it's the best and longest attested branch of the family". Can you imagine an IE specialist just arbitrarily deciding to ignore Celtic? But I digress...

I do have access to a fairly large Semiticist library; if you want to give me some time, I can do some checking and get back to you. Believe me when I tell you it'll be fun for me too.

The construct in different languages (and PS) also interests me a bit now, after looking into how Arabic forms it. When the head (construct) noun is singular, Hebrew and Arabic are quite similar - masculine nouns are endingless and indefinite, and feminine nouns regain the -t that they had lost elsewhere. When the head noun is plural, though, the two languages go completely different directions:

What you are saying is broadly true of spoken Arabic. In CA, however, the situtation is more complicated. In CA nouns lose their تنوين /tanwīnun/ "nunation" (assuming it's منصرف /munṣarifun/ "fully declined) in the construct state, but otherwise maintain their case endings:

x مديرٌ /mudīrun/ "manager" > مديرُ الشركةِ /mudīru-š-šarikiti/ "The manager of the company"
x في سيَّارةٌ /fī sayyāratin/ "in a car" > في سيَّارةُ المديرِ /fī sayyārati-l-mudīri "the manager's car"

The deletion of نَ /na/ (نِ /ni/ in the dual) at the end of masculine external plurals is probably a similar phenomenon. The two most common theories I've seen about this is that the *-n/-*m ending is a trace of an older, fossilized article (some say definite, others say indefinite), or that it was originally a masculine ending that later spread to the feminine for some unknown reason. To make matters more confusing, there are some OSA languages that used -n for definite nouns and -m for indefinite ones.

The construct noun being unable to take an article is pretty typical across the family. Aramaic was pretty similar to Arabic and Hebrew (except its article was post-positioned): *šam "name" *šamā "the name" "*šma malkā "the name of the king" (vowels may be off somewhat). What's really interesting to me is the way Ethiopian languages handle genitives; that is even though they've all pretty much become left-branching, they still retain this "construct noun can't be definite" rule. So you end up with things like: *yä.täbət.i bet (of.boy.the house) "the boy's house" - the opposite order of Arabic and Hebrew, but similar grammar otherwise. Clearly, this "construct" mechanism is a very integral part of Semitic grammar.

So far as I know, Semitic languages that didn't mark definiteness had different ways of marking the construct: Akkadian just lopped off the case ending: ālum "city" > āl šarrāqī "city of thieves", whereas Ge'ez used the accusative: bet "house" > beta nəguś "king's house".

Now the retention of the feminine -t in the construct isn't just a Semitic thing, but an Afro-Asiatic one. Egyptian did it (right up until Coptic), and there are traces of it in other branches (Berber springs to mind). Unfortunately, I can't think of any examples.

Did Proto-Semitic form construct plurals by just lengthening the case ending vowel of the singular, then?

Well, there is some evidence from the rest of AA that there was a plural marker -w which later elided into the long vowels present in ie. Arabic and Akkadian. Some very common nouns actually have a plural element -h- as well. Either one could explain vowel lengthening. Off hand, I can tell you that Egyptian plurals ended with -w (m) and -wt (f), possibly *-awa/-uwa and *-awat (bear in mind that our understanding of Ancient Egyptian phonology is highly conjectural and mostly based off of Coptic, spoken several thousand years later). However, Berber nouns usually have a masculine plural ending -īn/-ēn, so the so-called "nunation/mimation" present in Semitic probably dates back to AA in some capacity.

[Mecislau]

Hey, sorry for the delay. School's started again, so I haven't got a great deal of time.

Apparently, Aramaic used to "randomly" alternate between /C:/ and /nC/ in such verb forms as well.

In intensives? Really? Wouldn't that require the /n/ to have been preserved in Proto-Semitic to have survived in Aramaic? If so, why is there no other evidence of it in Semitic languages (or is there?)

If it was an independent development, that sure is a weird one - all geminates start spitting out /n/s.

I do have access to a fairly large Semiticist library; if you want to give me some time, I can do some checking and get back to you. Believe me when I tell you it'll be fun for me too.

Awesome. If you can find something, I'd be very grateful.

The University of Maryland library (the closest to my house) has a huge linguistics section, but is sadly lacking in information on early Semitic languages...



The Arabic nunation marking the indefinite - was that an Arabic innovation, or is it older?

And as I understand it, the Arabic broken plurals largely came from earlier collective forms. Has there been any attempt at categorizing all of these collectives in Proto-Semitic? Since there's such a large number of broken plural patterns in Arabic, I'd imagine that at least some of the original patterns must have had specific meanings.


Does Arabic (as well as any other Semitic languages) have a distinct set of endings for plural possessives (ie, "my dogs" vs "my dog")? Hebrew has a set of singulative and plural endings for this, but originally the plurals seem to have come from an infix -ay- inserted between the noun stem and the singulative possessive endings (eg, klavay "my dogs" < *kalb-ay-iya, or something like that). Is this *-ay- element seen in other Semitic languages as well? Do you know what that *-ay- actually is?

[hwhatting]

Granted, both are pretty rusty for her, and she's not a native German speaker (for that matter, she's not a native Latin speaker either. ).

So she's not from Latin America? (I know, old joke.)

Marcum Lucium interfecisse dicunt.
Lucium interfectum esse dicunt.

Interesting (bear in mind I haven't studied Latin in almost ten years now). Tell me, is this construction valid for any subordinate clause, or does it only apply to, say, "indirect reporting" verbs? What about existential/inchaotive constructions? For example, in Arabic, you can say:

من الممكن أنّ سليماً قتل حسناً
mina-l-mumkin.i 'anna salīm.an qatal.a ħasan.an
from-the-possible.gen that Saleem.acc killed.3ms Hassan.acc
"It's possible that Saleem killed Hassan"

لا شكّ من أنّ سليماً مجنونٌ
lā šakk.a min 'anna salīm.an majnūn.un
no doubt.acc from that Saleem.acc crazy.nom
"There is no doubt that Saleem is crazy" (Obviously because he killed poor Hassan )

In Latin, AcI is possible after verbs of reporting as well as after verbs expressing opinions and value judgements (like oportet "ought to" or licet "it is allowed").

I'm not sure that it would be used in your second example, but the first would be o.k.:
Non dubitatur Marcum dementire. "It isn't doubted that Marcus is crazy." NB that in Latin, AcI is used after verbs, not after nominal constructions like the ones in both of your Arabic examples.

I'm actually quite surprised neither you nor Maknas found أتى سليمٌ حسناً بكتابين "Saleem brought two books to Hassan" strange as I've had people tell me that "came X Y with Z" confuse them. However, my wife thought it was merely a double accusative verb, as the dual only has nominative and oblique forms. She suggested I give a less ambiguous example:

أتى بكتابٍ سليمٌ حسناً
'atā.0 bi.kitāb.in salīm.un ħasan.an
came.1ms with.book.gen Saleem.nom Hassan.acc
Saleem brought a book to Hassan.

I have seen directional accusatives too often to be surprised by that part, but I would have probably guessed for hours how Saleem came "in a book", as I tend to forget that "in" is only one, and not the only, meaning of bi.

As I said, it's been a while since I studied Latin, but even I can remember constructions like noctem totem dormivi. No, I couldn't remember the Latin word for Morning, sue me.

The word is mane (neuter). And I won't sue you, but correct you (correcting is more fun) - it's noctem totam.

Does Arabic (as well as any other Semitic languages) have a distinct set of endings for plural possessives (ie, "my dogs" vs "my dog")?

I assume Wiseblood comes back with a much more detailed answer, but based on my little knowledge, the answer is no for Literary Arabic (MSA). The same possessive endings are attached to singular and plural forms of the word. That said, there are some variants of the 1st p. sg. ending, but they are not triggered by sg. vs. pl.; AFAIK they depend on the phonological shape of the word.

[Wiseblood]

Hey, sorry for the delay. School's started again, so I haven't got a great deal of time.

No worries,as I'm quite late getting back to you now. I'm actually just now returning to school after three or four years away. I think I'd actually forgotten what a bump and grind it could be.

How's the Arabic going?

In intensives? Really? Wouldn't that require the /n/ to have been preserved in Proto-Semitic to have survived in Aramaic? If so, why is there no other evidence of it in Semitic languages (or is there?)

Well, the evidence is actually pretty scant: East Semitic had a "frequentative/iterative" form with a -tan- infix. Also, Arabic verb forms like افعنلل /ifʕanlala/ and افعنلى /ifʕanlā/ are supposed to be related as well (although note that the /n/ is after and not before the 2nd radical).

I've also seen the theory that the nasal infixation was a form of dissimilation that got generalized. I don't know any historical examples off the top of my head, but even today there are Arabic dialects that do similar things - in North Africa, their word for earthquake is something like znzal, from CA زلزال /zilzālun/. I've also heard people say طنطب /ṭanṭab/ for طبطب /ṭabṭab/.

If it was an independent development, that sure is a weird one - all geminates start spitting out /n/s.

Well, we can (roughly) divide Semitic langs into two types, those that mimate/perform nasal assimilation and those that nunate/don't perform nasal assimilation. This probably reflects dialectual variation within PS itself (or subtrate influence, take your pick), as both types can appear in most branches of the family. As Aramaic speakers were nunaters oftentimes surrounded by mimaters, random nasal insertion in geminates may have been a form of hypercorrection; that is given Caananite forms like *'atta for Aramaic *'anta, they may have added nasals to instensives to make the more "Aramaen" as opposed to "Caananite".

This isn't totally set in stone, as most Semitic languages reflect both types - as a nunating language, Arabic is mostly non-assimilating, there are some exceptions however, mostly in Qur'anic recitation: لن يذهبو Qur'anic /lay-yaðhabū/ versus MSA /lan-yaðhabū/ "they won't go"; Eblaite was a mimating language and did nasal assimilation everywhere, except in its pronouns.

As a little side note while we're talking about Aramaic and nasals: one of the arguements for grouping Caananite and Arabic against Aramaic is that Armaic has no N-stem verbs.

Awesome. If you can find something, I'd be very grateful.

Well, I've made a pretty thorough search of the library, as well as done some searching some academic sources, and even spoke with the local comparative Semitic guy - the conclusion I've come to is that there doesn't appear to be any kind of definitive reference work on Proto-semitic. Interestingly enough, a lot of the work being done right now is in Russian (a lot of it seems to be tied to Nostratic, though).

The University of Maryland library (the closest to my house) has a huge linguistics section, but is sadly lacking in information on early Semitic languages...

I'm very fortunate to be attending Memorial University, which (so as I'm aware) has the largest academic Semitic/Afro-Asiatic collection in Canada east of Ontario.

The Arabic nunation marking the indefinite - was that an Arabic innovation, or is it older?

This is a common misconception, held even by some Arabic grammarians. Nunation doesn't mark the indefinite per se, but rather the lack of a definite article/the absolute state in certain nouns. There actually are several kinds of awzan that don't nunate, irrespective of definiteness/state. Actually, one of the cool things about Arabic is that you can usually predict which declention a noun belongs to based on its wazn. Also, nunation isn't deleted in the sound masculine plural/the dual, unless the noun is in the construct state. Most likely, the deletion of nunation in definite Arabic nouns is a phonetic change brought about by the addition of an extra unstressed syllable to a word.

And as I understand it, the Arabic broken plurals largely came from earlier collective forms. Has there been any attempt at categorizing all of these collectives in Proto-Semitic? Since there's such a large number of broken plural patterns in Arabic, I'd imagine that at least some of the original patterns must have had specific meanings.

Heh, this is something I'd actually like to research myself someday. The broken plural phenomenon almost certainly dates back to Proto-Semitic, and there's a good chance it dates back to Proto-Afro-Asiatic as well (based on evidence from Berber and Beja anyway). Part of the problem with determining exactly what dates back to PS is difficult though, as most of the non Arabic/South Semitic evidence comes from early (and unvowelled) Aramaic texts, as well as really archaic East Semitic stuff, most especially from Ebla (which is bloody difficult to analyze on a phonetic/phonemic level, due to how rubbish cuneiform was at representing Semitic languages).

Anyway, I don't know if I've ever seen any kind of research trying to assign specific meaning to any of the broken plural forms. In all honesty, I'm not quite certain if this is the right way to analyze them myself, but I admit that it is an interesting idea. Part of the problem in doing this is that, while there are a few noun forms specific to broken plurals, most of them look just like nouns in the singular.

Be warned: This is going to get ranty.

Much like PIE, PAA was probably an negative language (Modern Berber and some of Cushitic are ergative, Ancient Egyptian looks a lot like an ergative language) with "animate" and "inanimate" genders (Actually, they may have divided them further than that - one of the explanations for the number of 3-radical roots with semantically related meanings that have the same first 2 radicals is that the so-called "3rd radical" was some kind of gender affix). Clearly, the animate gender constituted "living" things, while the inanimate covered non-living/abstract ideas (although what "living" and "non-living" meant to a PAA speaker is open to debate). Inanimate nouns probably couldn't be the subject of transitive verbs and probably couldn't pluralize either (they could be mass nouns, however). Now, for whatever reason, PAA speakers decided that they wanted to use inanimate nouns as transitive subjects, but they needed a way to mark them - basically, the *-at suffix that marks feminine nouns probably didn't originally mark the feminine per se, but was probably some kind of suffix enabling "abstract" ideas to become the subject of the verb. It probably later came to be associate with the feminine due to animacy hierarchy shifts (males were "more animate" than "females"). This would explain why a large number of "primitive" feminine nouns like *'ummum and *ʕaynum don't have the suffix. I've actually seen arguments that this ending is related to the *ta- prefix marking reflexives/passives, as well as the Ancient Egyptian impersonal pronoun tw (whatever it's original function, it was clearly some kind of "valency marker").

Now, I don't know how clear that was (or will be), but basically, this would explain why all inanimate plurals in Arabic are grammatically feminine singular. It also apparently explains some of the strangeness you see in Semitic accusatives with respect to subject/predicate distinctions. Coming 'round-about on this, in PS, you could pluralize "leftover" inanimates of the feminine gender through stem alternation, or by deleting what had by this time become the "singulative" *-at suffix. In some cases, this probably brought about stress change, further altering the "singular" and "feminine" forms. The broken plural system we see in Arabic/South Semitic today is probably the result of extending this feature to "male" and "animate female" collectives as well.

I think I may have just confused myself.

Does Arabic (as well as any other Semitic languages) have a distinct set of endings for plural possessives (ie, "my dogs" vs "my dog")? Hebrew has a set of singulative and plural endings for this, but originally the plurals seem to have come from an infix -ay- inserted between the noun stem and the singulative possessive endings (eg, klavay "my dogs" < *kalb-ay-iya, or something like that). Is this *-ay- element seen in other Semitic languages as well? Do you know what that *-ay- actually is?

I assume Wiseblood comes back with a much more detailed answer, but based on my little knowledge, the answer is no for Literary Arabic (MSA). The same possessive endings are attached to singular and plural forms of the word. That said, there are some variants of the 1st p. sg. ending, but they are not triggered by sg. vs. pl.; AFAIK they depend on the phonological shape of the word.

Hwhatting is right: Arabic suffix pronouns change based on the preceding word's "shape", not on morphological properties. For example, -ī becomes -ya after long vowels: عليّ /ʕalay-ya/ "on me", مستشفاي /mustašfā-ya/ "my hospital", مهندسيي /muhandisī-ya/ "my engineers". Actually, I kind of suspect that these "plural possessives" are really nothing more than the result of similar forms in Hebrew going through the crazy Caananite vowel shift. Do you think you could give me more examples?

So she's not from Latin America? (I know, old joke.)

I miss Dan Quayle.

In Latin, AcI is possible after verbs of reporting as well as after verbs expressing opinions and value judgements (like oportet "ought to" or licet "it is allowed").

I'm not sure that it would be used in your second example, but the first would be o.k.:
Non dubitatur Marcum dementire. "It isn't doubted that Marcus is crazy." NB that in Latin, AcI is used after verbs, not after nominal constructions like the ones in both of your Arabic examples.

I think I can bring this discussion to a crashing halt, at least in terms of the "exoticness" Latin AcI: We do the same thing in English! "I think him to be crazy." or "I've known her to visit the store."

I have seen directional accusatives too often to be surprised by that part, but I would have probably guessed for hours how Saleem came "in a book", as I tend to forget that "in" is only one, and not the only, meaning of bi.

You know, I almost never associate "bi" with "in", but rather "at" or "by/with". I've always thought "fī" better translates "in".

The word is mane (neuter). And I won't sue you, but correct you (correcting is more fun) - it's noctem totam.

See above about Dan Quayle.