Grammar in Arabic education

In July 2017 I inquired via the Linguist List about the teaching of grammar in the Arabic-speaking world and its historical links to the medieval Arabic grammarians. I received very helpful answers from: Alexander Magidow, Enam Al-Wer, Clive Holes and Jonathan Owens. Here they are, lightly edited and in chronological order:

Alexander Magidow (1)

I saw your post on Linguist List. In the Arab world, they don’t actually learn any graphic system as far as I’m aware, but there is an extremely complex system of grammatical analysis based broadly on the works of classical Arabic grammarians dating from the 8th century or so. Most students are completely terrified of it, and to a large degree it seems to primarily serve to alienate students from their languages (See Haeri’s “Sacred Language, Ordinary People”). However, students are expected to be able to identify parts of speech, parts of sentences (subject, predicate, verb, subject of verb, object of verb), with most of the focus on how words’ case endings change based on their place in the sentence or on governing particles. In some places, there is a standardized way of expressing these terms, but it’s prose, e.g. “This word is a noun functioning as the subject of a verb and therefore it is marked as nominative with the marking of the case taking the form of a ‘u’ vowel.”

I’ve seen some reference works that attempt to provide color coded bubbles to distinguish cases, but other than that it’s largely a prose, not graphical, system. It is quite complex though, and recursive – you can have a sentence operating as a predicate, for example, or even as a direct object, in the system. It is similar in some ways to American generativist analysis, especially if you read the old grammarians who are concerned with a form of governance (in a manner very reminiscent of G&B stuff.)

Since it’s not a graphical system, showing you examples from textbooks probably wouldn’t be too useful.

Alexander Magidow (2)

In the Arab world, teaching of the Classical Arabic (well, potentially modernized, though I’m actually doing research on how exactly the language changed and when) is standard in schools, and the ideologies of language are such that Arabic=Classical Arabic, and modern dialects are treated as unfortunate deviations from that norm. Students in the appropriate specializations are tested pretty rigorously on it on their end of high school baccalaureate-style exams.

What I was taught was done mostly orally via a very, very old school Arabist. In his method of analysis, you CAN analyze a clause in the same way as a single word, and I have some reference grammars (in Arabic, for Arabs, I think for helping with school exams) that use a similar approach, so it is a way of teaching grammar used in some schools I think. Normally the clause must have an appropriate nominalizer, e.g., with three lines, text, normal gloss, grammatical analysis. Excuse the kind of ad hoc nature of it :

ʾu-rīd-u tufāḥat-an
I-want-indicative apple-acc
Implicit.subject-verb object
“I want an apple”

ʾu-rīd-u [ʾan ʾa-ḏhab-a ʾilā l-sūq-i]
I-want-indicative nominalizer I-go-“subjunctive” to def-market-genitive
Implicit.subject-verb object=verbal sentence[particle Implicit-subject-verb preposition noun]
“I want to go to the market”

You can also get a nominal (e.g. equational sentence, since no verb to be is required in present tense) sentence analyzed as “subject” and “predicate” but you can have a verbal sentence as predicate (this happens most often if you start the sentence with a pronoun – since Arabic is a pro-drop language, this is considered poor style, but happens), OR you can have another nominal sentence as predicate:

ʾana ʾu-rīd-u tufāḥat-an
I I-want-indicative apple-acc
subject predicate=verbal sentence[Implicit.subject-verb object]
“I want an apple”

Zayd kitāb-u-hu kabīr-un
Zayd book-nom-his big-nom
nominal sentence=[subject predicate=nominal sentence[subject predicate]]
“Zayd, his book is big”

This is a kind of basic analysis – as you said, it’s normally expressed as parsing, going word-by-word. I’m just not 100% that most school children have to do parsing at the phrase level, but it definitely is done at some level.

[me] Many thanks once again for a fascinating message. What you say sounds rather like what I read about in Owens, and I’m just wondering whether the same system has been perpetuated unbroken since the 8th or 9th century! It would be wonderful if you could make inquiries about current practice, and meanwhile I know one or two people I could ask.

Well, I doubt anything is truly ‘unbroken’ but the basics of grammatical analysis in Arabic hasn’t changed that much since Sibawayhi – it’s one reason he’s so highly regarded, and his book, “Al-kitaab” is one of two books you can call “al-kitaab” and have people understand which book you’re referring to. The main reason for the persistence of the grammatical system is that it was meant to explain the Quranic case system (the other early grammatical work is “Meanings of the Quran”), and early/pre-Islamic poetry. Since the corpus hasn’t changed that much, and since you’re not supposed to apply it to more modern forms of Arabic (since they’re just ‘bad Arabic’), there’s no reason to modify the system that much. That, and the quality of his analysis is pretty high quality – a lot of later stuff is quibbles about theory internal stuff, though I could be minimizing the differences there since it’s not my area (but there’s a decent amount written about it.)

There were of course various different pedagogical attempts – probably the most famous is the ‘Thousand’ by Ibn Malik (d. 1274), who composed a 1000 line poem that encapsulates the grammatical system for pedagogical purposes. I knew someone who actually played a recitation of it in his car. I don’t know how much it’s actually used today though. Apparently there was a similar work, the Ajaaramiya, which is available online in translation, that might give you a sense of what pre-modern, and modern but old school instruction is like. I don’t see them talking about nesting here (though see page 12, where you do have complex predicates, meaning more than just a bare nominal/adjective though the examples aren’t translated), but that may fall by the way-side in some styles since a whole phrase doesn’t receive any case marking, and therefore is somewhat irrelevant if your only task is figuring out the appropriate case endings:

I’ll ask some colleagues what their school instruction actually entailed in terms of analysis.

[me] Can I just check the bit about subjects and predicates, please? The Latin grammarians assumed, following the Greeks, that every sentence has a subject and predicate, but Arabic grammarians seem to have restricted these terms to ‘nominal’ sentences like muhammadun rajul-un (or whatever – I know a bit of spoken Sudanese Arabic, from when I did fieldwork on Beja – Cushitic). But if there’s a verb, the subject is just a dependent like the object? Is that right?

There is a distinct difference in Arabic between a nominal and a verbal sentence in the version of Arabic grammar I’ve seen, determined by the POS of the first content word in the sentence. So a nominal sentence consists of a subject and predicate, and the verbal sentence consists of a verb, subject (which may simply be ‘hidden’ in the verb) and potentially a direct object. As to whether those are considered ‘dependents’ probably requires some time looking at the terminology. And of course, the terms ‘subject’ and ‘predicate’ are a bit different in their uses probably.

There is a concept of meaningful sentence parts that crosses both types of sentence. You have the concept of ‘isnaad’, the juxtaposition of two things which together complete a sentence, with “al-musnad”, which in the verbal sentence is the verb, and in the nominal sentence is the predicate, and “al-musnad ilayh” which is the subject of the verb or the subject of the verb.

[me] Thanks. What you say sounds as though the verb and the predicative nominal are both considered heads, but (so far as my limited knowledge of Arabic grammar goes) the verb follows the regular head-first order while the nominal is second, so it’s not obvious that it’s the head.

Enam Al-wer

[who went to school in Jordan]

As far as I know nothing has changed since my days, except to include more Quranic verses. It starts with ‘language/speech is divided into nouns, verbs and particles’, then the sentence. Parsing is introduced very early on. The theoretical framework is GB-like. The focus is on syntax, then morphology; a bit of phonology, including assimilation.

[commenting on my summary bullet points]

  • It’s only classical Quranic Arabic that’s taught; modern varieties are ignored as corrupted variants. 

Definitely. Any modern construction is considered wrong if it contradicts the prescriptive rules laid out by the ancient grammarians, even if it’s current in Modern Standard Arabic and used widely in the written media. One example that sticks in my mind comes from a meeting I attended at the language academy in Jordan where there was an objection to the use of the verb ‘raʃad’ as a transitive verb – not sure how to translate it, something like ‘to know, acquire knowledge, grow in knowledge’. In that year, 1980s, stickers were being used next to light switches with the expression ‘tarʃi:d il ʔistihla:k’ (ration/limit consumption), meant to remind people to switch off lights to reduce consumption of electricity. In this expression the derivative form ‘tarʃi:d’ is followed by an object. They spent most of the session discussing this catastrophe, and decided to file a complaint demanding that the stickers be withdrawn. They were never withdrawn and the expression is widely used.

  • The syntactic model is based on word-word dependencies, which can include subordinate clauses introduced by a subordinator.


  • The morphology is frighteningly complicated.

Horrendous. It took morphology and me 30 years to reconcile.

  • It’s very boring.

You can say that again and again …. Add, no one remembers anything at the end of it, only the anxiety remains.

In my days, private schools in Jordan were not allowed to teach the sciences through the medium of English only; so even those pupils who wanted to go for GCE/A-Level instead of the Jordanian school leaving examination had to cover the curricula in Arabic too. This regulation was abolished some 20 years ago; the vast majority of the kids I know opt for the English system or IB to avoid as much Arabic as they can – but they still have to do Arabic as a subject. The teaching of the grammar is made worse by the archaic language of the examples used to demonstrate the various constructions – many come from the Quran or old poetry, incomprehensible. I understand the Lebanese school books are different in this aspect.

Clive Holes

To add to what Enam says: I have long been of the opinion that the archaic materials and antediluvian methods of teaching Classical Arabic — exactly as if it had not moved on since the 8th century — are a potent symbol of the sclerosis which has gripped teaching and learning in the Arab countries at large, and, frankly, a token of the inability to think creatively. Any kid with a brain runs a mile from it and embraces western modes of thought and analysis. Sounds a profoundly non-PC, even racist thing to say, but that’s the brutal truth. The most illiberal, hide-bound, inward-looking people I ever met in the decades I spent living and working in the Arab World were the ones who spent their time inculcating, usually with the help of a stick which they had no hesitation in using, a non-negotiable set of incomprehensible grammatical rules into classes of innocent young school-children.

[me] Wow! And that’s clear too.  And no doubt you might even add that this kind of mindset sits well with Islamic extremism. One thing I’ve found in my research is that grammar can be surprisingly close to politics; e.g. Henry VIII used grammar as a tool for getting power off the church. This looks to me like another case of that. It must be so frustrating being an open-minded intelligent youth in the Arab world. Does anyone know if this is an Arab thing or a Muslim thing? E.g what happens in Iran?

Well, what I do know to be the case is this: on the rare occasions when, before the English voice-over kicks in, one hears representatives of ISIS, Hamas and Hizbollah talking on camera in the western media, they invariably speak (or attempt to) in Classical Arabic with full mood and case endings, whatever they are talking about. It sounds faintly ridiculous to ordinary people, but here the medium is most definitely a key part of the message. The word 9arabiyy ‘Arab, Arabic’ occurs a dozen times in the Qur’an, and in every case in collocations like lisaan-un 9arabiyy-un mubiin ‘a clear Arabic tongue’ when it is telling us that the Qur’an is in a language that ordinary people of the 7th century could understand (unlike Greek, Aramaic, Latin, etc). That language came to be known as ‘Classical Arabic’. So every religious zealot with an agenda will automatically use it, even in the 21st century. Use of it is loaded with symbolism.

Most Muslims outside the Arab World, unless they have special training, don’t  understand any kind of Arabic except for odd religious fixed phrases — it is certainly not a functional language for them. A bit like Latin was in the Catholic Church for the average Catholic.

And a postscript: one of the big experts on the system of description the Arabs developed to codify and explain the rules of Classical Arabic, Mike Carter, wrote his PhD thesis (recently published after about 40 years) on how Arabic grammar was based on an extension of the principles of legal argumentation (which in the mediaeval Arab World meant the Shari’a, which was then the only game in town). The same terms key terms are used, e.g. qiyaas ‘analogical reasoning’.

What ISIS, Hamas, Hizbollah do in terms of language choice is just that, a conscious choice, it isn’t inevitable; there are well-known TV imams in Egypt, for example, who will quote a piece of scripture with faultless Classical Arabic inflections and then go on to explain what it means in the everyday Egyptian colloquial of their listeners. That is how it used to be. They didn’t need to parade their language credentials to show how pious they were like these modern zealots.

Enam Al-wer

I totally agree with Clive.

I have a theory, some may think wishful thinking, but here you go: the insistence on the part of the fundamentalists on using classical Arabic will deal the last blow to the language. There are strong signs that it is increasingly being associated with criminal acts, and everything unattractive. The situation reminds me of the demise of Katherevousa; in one analysis, can’t remember whose, it is thought its use by the Junta and the conservative forces, while progressive thinkers used Dhimotiki, established a link between K and conservatism in politics, and support for the status quo during the height of the struggle.

[more questions from me]

  • Teachers’ subject knowledge: do teachers study grammar at university before going back into the classroom as teachers?

Yes, normally.

  • Did I ask about diagramming? If I didn’t, I meant to ask you whether there’s any kind of diagramming system for syntactic structure, e.g. trees, tables, indented clauses, etc. 

No, don’t remember there being any.


Jonathan Owens

[My questions are in italics, and are based on earlier discussions with Alexander and Clive.]

Q.1. Is the conceptual framework still the dependency-based system of the classical grammarians? I’m particularly interested in the place of the subject-predicate distinction, which seems to me to have prevented western grammarians from developing a consistently dependency-based theory (because clauses were always different, and had no head) – until Tesniere sorted it out. Clive and Alex tell me that Arabic grammar recognises ‘subject’ and ‘predicate’, but I think they’re saying that these terms are reserved for nominal sentences with an understood copula (‘Muhammad man big’). Would they recognise a head in such a sentence?

One of the highpoints of the Arabic tradition is the discussion about what the governor of a (morphological) verbless sentence is. The entire discussion was encapsulated in a brilliant chapter by the mid 12 century grammarian Al-Anbari who argued that although a nominal sentence is verbless, it does have a governor. A sentence like  zayd-un mudarris-un ‘Zayd is a teacher’ would have the structure as given in the attachment [not included here]. Essentially what he argued was that the simple indicative, non-topicalized, present tense nominal sentence has essentially the same structure as related sentences, for instance one with a past tense verb (kaana) or with a topicalizer inna. He observed that since both of these sentences have an overt governor, it can be argued that the formally unmarked (by a governor) present tense sentence in fact has an abstract governor. Some grammarians said that the “fact of beginning” was the governor. Anbari’s formulation with an abstract governor I think is the best analysis. The argument in favor of this abstract (essentially a 0 governor) governor was that in the related past and topicalized sentences the two dependent nouns change their case ending according to governor. In the case of kaana ‘was’, the subject is nominative, the complement accusative, while with inna it is the opposite, the subject is accusative and the complement nominative. So the reasoning was, since the subject-predicate relationship is essentially the same in all three cases, and given that in two cases a simple dependency relationship is universally accepted (in the Arabic tradition), it is a small step to assume that there is a governor in the case of the unmarked simple present-tense sentence. Clearly a motivation here is the desire to create a uniform structure for all sentence types in Arabic, essentially successful as I see it. I discuss all this in detail on p. 55 in my 1988 “Foundations of grammar: an introduction to medieval Arabic grammatical theory”, along with related but more complicated structures. I also lay out the argument in that chapter for understanding the Arabic thinking as a dependency grammar.

Q.2.Has anyone ever suggested any kind of formal notation for dependency analysis such as lines/arrows or a table?

So far as I know, there was never a visual representation of dependency in the Arabic tradition. In general Arab-Islamic culture wasn’t  much into stylized representations of structures.

Q.3.Is Arabic grammar inevitably the grammar of Quranic Arabic, rather than of a modern variety? I.e. is it always a vehicle for teaching a dead language (like Latin in post-Roman Europe)?

First, it is a major misconception that Modern Standard or even Classical Arabic is the same as Quranic grammar. It does have to be said they are very closely related. But myself, having wrestled with an historical interpretation of all the material , I think one has to start with Classical Arabic from about 920 CE. This is about 300 years after Muhammad, and over 100 years after Sibawaih, who pretty much laid down everything we know about Classical Arabic. What happened about 920 is that a grammarian named Ibn Al-Sarraj wrote an exquisitely compact and readable grammar of (now we can call it) Classical Arabic, which pretty much set the standard for everything since, including Modern Standard Arabic. So yes, Modern Standard Arabic is inevitably a grammar based on a classical source, but not on the Quran (rather, for a number of interesting reasons, the Quran needed classical Arabic before it could assume its standardized form).

A very interesting historical linguistic issue is that the Quran itself did not become standardized until the same era (930, Ibn Mujahid). Why not? That is the interesting historical issue. There are a lot of mythologies developed both by Arabs and by western Semiticists which will tell you what happened linguistically, though since this is a matter I haven’t thought through completely myself I don’t want to commit myself to a particular position. I am pretty sure that from an historical linguistic perspective (which Arabicists and Semiticists hardly ever assume as a starting point) the matter is many facetted, complex, and more interesting than the simple mythologies you will get from most people who you pose the question to.

Of course, people are always trying to “modernize” the classical language; there are Arabic institutes all over the Arab world, though what is largely understood here is the creation of vocabulary to deal with modern terminology. There is little discussion of modernizing grammar, and indeed it is hard to see how this would happen. This has the result that Modern Standard Arabic is pretty much alive as a written language, but rather dead as a spoken language.

Q.4. They also tell me that Arabic grammar, as taught, is boring and incomprehensible. If so, why? Is it because it’s seen as in some sense divinely inspired, like the language, and therefore not to be tampered with?

This sort of leads to the last point. I have to say that language teaching never was a particular interest of mine. More a necessary evil. If people want to learn a language they will. Moreover effective lg teaching is maybe 70% teacher, 30% material? That said, I think two factors in the West militate against an effective Arabic teaching program. The first is the reality of Arabic. Classical Arabic/Modern Standard Arabic is a written language and all the gimmicks and tricks one might want to think of to make it into a normal spoken language are pretty useless. Arabs don’t speak it, so why should westerners? 30 years ago this wasn’t a problem. No one cared about learning to speak Arabic. Universities taught classical Arabic, students learned it. I suppose everyone was happy. I didn’t study Arabic at SOAS, but I do remember taking a course in an Arabic dialect from T. M. Johnston, whose work on Gulf Arabic is read until today. As a course, however, it was pretty hopeless. He assumed one knew classical Arabic, and so he could run through the colloquial grammar at breakneck speed. I didn’t finish the course with him, which mattered to no one. But now, quite reasonably, people do want to learn to speak, and the only spoken variety is the dialect, one or another. This point has now been recognized in most Arabic teaching programs, though here there are two constraints. One is the cultural prejudice which still obtains against the dialect, so there really aren’t particularly good teaching textbooks. The bigger constraint, however, is time. Universities really don’t give enough hours so students can properly learn both classical Arabic and a dialect. Choices have to be made.

That’s the situation in the West. I don’t know the Arab world all that well, though what I know of it is not pretty. The only Arabic is classical Arabic and the rote method is the favored methodology. Memorization. So long as enough time is devoted to it, the method works I think, but it requires time and motivated students and my guess is that it is pretty inefficient and ineffective against the goals of language learning and the amount of time invested. However, this is simply another research niche which never gets any attention in the Arab world itself (and as far as I’m concerned, there are more interesting things to deal with as far as Arabic linguistics goes). In this regard, I  have given lectures at universities/conference in different Arabic countries, in which I discuss issues like Al-Anbari summarized above, with the view to saying, look, Arabic linguistics is really quite interesting, intellectually challenging and very compatible with aspects of contemporary linguistics. I’ve given talks both in Arabic and in English, the result being the same in both cases. Oh. So what. If I’ve given the talk in Arabic, praise for the language effort, but no reaction to the content.

Specifically in this case I blame Egypt, in particular the traditional methods of Al-Azhar and even of Cairo University. Graduates of those institutions have dominated Arabic departments throughout the Arab world, and have transmitted their anachronistic and myopic, Arabic closed-world view further, to Kuwait, Saudi Arabia, Qatar. The situation is basically hopeless. That at least would be my working hypothesis, if I were investigating the issue. Religion has a role to play here I suppose, but that doesn’t explain the close-mindedness relative to other linguistic ideas, or an ingrained lack of interest in the world outside of Arabic. To a large degree, moreover, it is a world view shared by much of the West, the Orientalists, the Middle Eastern Studiers, the Semiticists. Linguistics really doesn’t have a place in this world. Against this, at Bayreuth I’ve had ups and downs as far as the quality of students goes, but in the years when I had good students we would read the Arabic grammarians, the afore-mentioned Ibn Al-Sarraj in particular, and the students loved it. One student who had studied the same book in the traditional way in Damascus, i.e. listened to the professor read from the book and give a wooden explanation as to its contents, said that he didn’t understand the author until he took the course with me. Surprise? Hardly. The book is about grammar, and grammar has universal principles, and once you understand that, the rest comes easy. Unfortunately, as far as Arabic goes, it once came easier than it does today.



Comments are closed.

Set your Twitter account name in your settings to use the TwitterBar Section.