Thoughts on Arabic Diglossia from the Growing Participator Perspective
by Greg Thomson
Charles Ferguson (1959) coined the term diglossia to refer to situations where for centuries a community uses two substantially different but historically related language varieties for different purposes in their day-to-day lives. He mentioned four examples: Arabic, Swiss German, Haitian and Greek. There has come to be a large literature on diglossia in the past forty-five years, and the term has often been applied to situations in which the two language varieties in question do not differ from one another so greatly as in Ferguson’s examples.
A search in Linguistics and Language Behavior Abstracts combining the terms “diglossia” and “Arabic” turns up 110 citations to works published since 1993. It is not my purpose to delve into that literature. I leave that to some reader of this paper who is planning an M.A. thesis or Ph.D. dissertation, and would like to explore this matter more deeply as it relates to language learners.
Rather, I am reacting mainly to discussions I have had on the topic in the Arab world, with both Arabs and non-Arabs, in the light of comments by Badawi 2002. I must address this issue because of the way the diglossic situation interacts with the Growing Participator understanding of language learning. I find Badawi’s evidence credible in that it seems actually inconsistent with some of the conclusions he wishes to draw. I’ll begin with Badawi.
Badawi, following the standard convention, refers to the varieties of Arabic that children learn first as L (for “low”), and the variety that is learned through formal education as H (for “high”). On p. 158 Badawi tells us that
Modern H is fully functional as a written medium, although as a spoken one it has only narrowly prescribed formal functions which are poorly carried out by halting performers. Reading of news bulletins on radio and television, which is always carried out through H, cannot be considered as speaking. The situations in which spoken Modern H is used as a necessary medium of communication between individual Arabs are conferences and learned gatherings of Arab scholars from different Arab countries. Oral performances in H, which vary considerably in the level of proficiency from one individual to the next, cannot be described as native ability. (In fact, some, on such occasions, substitute the H with Educated Colloquial Arabic.)
On p. 159 Badawi adds that writing in L “parallels attempts to speak H”. That is people are clumsy and lack fluency (and orthographic consistency) when engaged in the uncommon practice of writing in L.
On pp. 158 Badawi talks about the linguistic distance that exists between different spoken Arabic varieties: “…intelligibility, particularly in contexts removed from learned topics, may become seriously impaired.”
In my experience not all Arabs emphasise the linguistic variety of spoken Arabics. There seem to be two reasons for this: pan-Arab nationalism and religious apologetics. The former may lead Arabs to assert that spoken varieties are similar to one another and to the H variety. In terms of religious apologetics, it is an important claim that the Koran is the only great holy book that is written in a current, living language. These two factors, political and religious, can make objective discussions of diglossia sensitive.
Other Arabs emphasise the distance between L (their native variety) and H (their educational variety). One woman told me that the early years of school were extremely difficult for her and her classmates because they couldn’t understand anything! She may have been exaggerating, but the point was clear that for her and her classmates it was not a small step from L to H. This would be consistent with the understanding of Western scholars, including Ferguson, whose very idea of diglossia required that the L and H varieties be essentially different languages, though historically connected.
According to the picture that Badawi paints, there is a major difference between Arabic diglossia and cases like Swiss and Haitian diglossia. In these latter cases, the H language actually has a community of native users. In the Swiss case, for example, that would be those Germans for whom “Standard High German” is the native language. In the case of Arabic, Badawi informs us that the H language has no “native speakers”. This has serious implications for understanding the nature of language processing in H Arabic, and also for the process of learning H Arabic. A language, in the common sense of the word, depends on the existence of production and comprehension mechanisms in the brains of members of the community which uses it. H Arabic is not a language in this normal sense. There may be a small proportion of Arabs who are relatively “fluent” in it, but they cannot reach the proficiency level that we in the English speaking world call a “Near Native” (FSI level 4) for the simple reason that there just aren’t any natives to become near to!
Badawi recognizes in a footnote on p. 159 that this fact is sometimes lost on Western specialists in Arabic as a foreign language:
…the ACTFL [proficiency level] guidelines [similar to the FSI proficiency levels] state that a “superior speaker of Arabic should have superior-level competence [roughly, FSI level 3] in both MSA and a spoken dialect and be able to switch between them on appropriate occasions”… To the best of my knowledge, no native speaker of Arabic has superior-level speaking ability in both.
“Superior” is a whole (very large) level lower than “Near Native” and two (very large) levels below “Native-Like” (FSI level 5). Thus, according to Badawi, not only are there no native speakers of H, but not even anything remotely approaching native-like speaking ability. It is important to understand what this means. One non-Arab argued adamantly with me that since both L and H are called “Arabic”, and since L has native speakers, then Arabs are also native speakers of H. It is not an issue of what a language or variety is called, but rather of the mental mechanisms that are involved in the comprehension and production of that language. The mental system of L, for Arabs, is that of a normal native language, while there are no Arabs who have a mental processing system for H that is like that of a normal native language, at least according to Badawi. In other words the mental processing system of H is fundamentally different in nature from the mental language processing system of L. This seems reasonable in view of the place that L and H have in Arabs’ life experience.
In the statement from the ACTFL guidelines, we see what I experience regularly in discussing Arabic diglossia with Westerners in the Arab world: they talk as though H and L are the same general kind of things: two dialects, two language varieties, maybe even two languages, but whatever L is, H is just another case of the same thing, let’s say, just a different dialect from L. Under such a presupposition, it makes perfect sense to debate whether one should learn L first, learn H first, or learn both L and H together. Yet from Badawi’s evidence we are led to believe that L is just a normal (preliterate) human language, learned from infancy, and supported by a normal auditory comprehension system and a normal speech production system. All normal natives are “native-like” in their L Arabic, and could potentially nurture non-natives more and more deeply into their native world. This is not the case with H. (We can assume that some system of fluent auditory comprehension mechanisms for H develops for many educated H users, but apparently not a normal enough comprehension system that it can support the normal development of a highly fluent production system.)
Educated Arabs are highly fluent in reading H. The reading process that is involved is presumably not the same as a reading process that is founded on a prior existing listening comprehension system. This is not to say that ability to understand L in no way contributes to “reading” ability in H. However, learning to read a language for which one already has a complete listening comprehension system is initially a matter of applying existing mechanisms to the written form. In general, when we read a language for which we cannot understand speech, it may be better to think of the “reading” process as rapid deciphering, rather than straightforward language processing like that involved in listening comprehension and in advanced L1 reading ability. With practice, people can become rapid decipherers. In any case, whatever it is that fluent readers do when they read H Arabic, it is not part of a full-blown language processing system.
What I am suggesting, then, is that L Arabics are garden-variety (preliterate) languages, while H is especially tied to a sort of “reading” process, and eventually, to the ability to understand material that is “read aloud”. This difference appears to be obscured for some foreigners learning Arabic, such as the authors of the ACTFL guidelines quoted above. I feel I have a good idea of why adult learners of Arabic do not experience a fundamental difference between learning L and learning H: almost all formal language teaching is “speech-led” and reliant on the written language (see Arthur, 1993). Language material is presented to students on the printed page (or board). Nowadays new material may also be pronounced aloud by the teacher, or accompanied by an audiorecording for students to listen to, but these audio forms are secondary. Students don’t complain if they are expected to work directly from the written form to the spoken form (following “pronunciation rules”), but they often become very anxious if asked to rely on the spoken form without the written form. So language learning, as popularly conceived, goes “from the page to the mouth” (with the possible assistance of a tape or native teacher).
By saying that an approach to language learning is speech-led, we mean that the language learners see themselves as exerting considerable effort in learning to speak, while they assume that less (or no) effort needs to be directed toward learning to understand speech. In the speech-led view of things, paradigms, grammar rules or model sentences are seen as guides for constructing sentences for the purpose of speaking. The written “rules” are like a recipe for speaking.
It seems clear to me that a language learner who follows a speech-led approach, and relies on written language in order to learn “how to speak” will have exactly parallel experiences in learning L and H. In either case it is a matter of starting with the written page, and developing the ability to make up sentences for speaking. There is not the sociocultural concept of native users nurturing us into deeper participation in their lives. There is the cognitivist concept of learning to make up sentences for speaking based on what one sees in writing.
Thus when I suggest that learning H and L are very different matters, many Westerners who have learned Arabic to a functional level seem baffled. For them learning the two has not been all that different. The vocabulary, sounds, and grammar forms are different, true, but the process of getting the language from the page to the mouth is the same. Don’t tell them it isn’t the same! They’ve done it! It’s the same! And I agree that for them, the experience of learning (and using) H and L could be highly parallel.
Another major complication involves the attitudes of Arabs toward “teaching” L. In fact, that may be nonsense. “Teaching” is only about H. The issue of interethnic/international self-presentation also comes into play. Arabs may be proud of H, feeling it to be rich, beautiful, in short, impressive to foreigners, while by contrast they may feel their own L to be an embarrassment. That is at least a common feature of diglossia. Thus the Arab social context presents great challenges to the GP approach, of the sort we commonly attempt to address. It needs to be recognized that learning L is more like learning a minority language than like learning a major world language, in that there will often not be the opportunities for constant exposure through the media. However there will often be some opportunities: radio phone-in discussion shows, T.V. talk shows, sitcoms. Written theatrical plays may even provide lengthy written texts in L, as may collections of jokes.
In the four-dimensional chocolate world of the Growing Participator Approach, learning L is a very different matter from learning H, although both processes are ultimately part of a single larger process of growing participation into an Arabic languaculture. The H-based cultural activities are not a failed attempt by Arabs to develop a second spoken language in addition to L. Rather, those activities have their own life and role that is profoundly different from the life and role of L, but intertwined with it in the web of life. Think of how it has been for Arabs: There are the L Arabics, which they learned to a normal level of a six or eight-year-old’s language ability, and which they then used massively as an aid to learning H. Then over time, there was the continuing impact of H on L, as they grow through various stages of educational, social and cognitive development, during their years of formal education and after. For certain well-educated people there comes to be “high registers of L”. These are varieties of L which have heavily borrowed from H. These are what Badawi calls “Educated Colloquial Arabic”. Educated Arabs with differing L Arabics will each, at an academic conference for example, speak their own local “Educated Colloquial Arabic” (which I’ll call Educated L) to increase the level of understanding over what is possible if each speaks the lower register of their own L Arabic, but allows much greater fluency than if they tried to speak to one another in H.
Now, we might ask, could a foreigner just begin by learning Educated L as their primary Arabic, since that is a register of a real living language that does in fact have proficient native users. Well, one can do many things, including speech-led, writing-dependent language learning. Certainly, many if not most adult language learners end up learning lots of “adult language” and missing lots of “four-year-old language”. From the Growing Participator perspective, we try to resist this tendency, to a reasonable extent. We believe that bypassing the language of everyday life in favour of the language of newspapers or academic books leaves us with low comprehension ability in everyday life—lacking thousands of basic vocabulary that we’ll never get around to learning. Of course, even with our good intentions, we end up compromising a lot, and in our first four years following the GP approach, we learn lots of language that four-year-olds don’t know, and miss lots that they do know. But we don’t make that our deliberate strategy. We view it more as an inevitable weakness rooted in our circumstances.
What, then, are the implications of Arabic diglossia for those attempting to follow the GP approach. We make the following tentative suggestions:
1) Avoid the fallacy that learning H and learning L are just two parallel instances of “language learning”, and thus that the only concern is deciding whether to learn L first, H first, or both simultaneously.
2) Recognise the enormity of the task of learning L, even without H.
3) Recognise that deep-life sharing is more of an L issue than an H issue. Apologetics with ordinary people may also be more of an L issue than an H issue.
4) Recognise the goals for H: be able to rapidly decipher written materials, be able to follow media Arabic and read contemporary literature (say, after three our four years of growing participation), and be able to read aloud with reasonable pronunciation.
5) Remember the temporal dimension. The ability to use appropriate styles and registers (H in H contexts, L in L contexts, Educated L in Educated L contexts) normally requires years of experience with the language (and is not an FSI level 3 ability as suggested by the ACTFL guidelines quoted above). Don’t be pressured to worry about using appropriate styles and registers during your early months—concentrate on being nurtured into L and if you wish, on learning to decipher written H.
6) If your social life is among educated Arabs and in academic contexts, then expect to eventually be nurtured all the way up into a full linguistic repertoire which includes proficiency in Educated L.
7) Depending on their role, some workers may want to develop reasonable Classical Arabic pronunciation in order to quote the Koran from memory and so on (not necessarily to be a serious reader of Classical Arabic); on the other hand, this may be of little relevance to you (say, if you mainly relate to uneducated women).
8) Don’t think of H as entirely an issue of private reading. Interacting around written texts in L or Educated L is an activity of the languaculture into which you are being nurtured. I probably would not do a lot of this until Phase 5, however, since by their very nature, written materials are “native-to-native” in terms of their difficulty level. Private reading (deciphering) might begin earlier, but probably not earlier than Phase 3.
Now many Westerners in the Arab world disagree deeply with the implications I have drawn from Badawi’s comments. This isn’t surprising if their basic approach to language learning is cognitivisist, speech-led and writing-dependent. They don’t experience H and L as being different kinds of things at all. In fact, they feel that H and L aren’t really all that different, although they usually admit that when they overhear native L users interacting animatedly, they cannot follow the conversation. They can follow the media language somewhat better, or maybe even quite well. But it is probably true that their foreigners’ version of “colloquial Arabic” has a higher ratio of “Modern Standard Arabic” as the host people’s “colloquial Arabic” does. If they cannot follow local speech, they may remain unaware of this fact, perhaps for years. The two factors conspiring to produce the result are 1) their speech-led, writing-dependent approach to language learning and 2) the host teacher’s discomfort with teaching L at all.
This course of events may give rise to new non-native Arabic varieties: International Business Arabic, Diplomat Arabic, Missionary Arabic, International Student Arabic, and so on. These foreigner Arabics could be disproportionately influenced by H. There may also be non-native Arabics such as Household Servant Arabic, Oilfield Labourer Arabic, Merchant Arabic, etc. These may have only limited H influence. Well, in our real interlanguacultural globalising world, we might consider the state of affairs of many “foreigner Arabics” to be normal, and healthy. I’m not against that idea. At the same time, I think the Growing Participator option should be available for those who want to grow into a rich new world in four-dimensional chocolate.
In conclusion, the answer to the question, “Should I first learn H, first learn L, or start learning both at the same time?” is “Why are you talking like that?” (That is, why are you equating learning H with learning L, as though it’s just a question about relative timing, and as though the goal is generally the same and the process the same.) Well, that is how I’d answer someone trained in the four dimensional chocolate GPA. To others my common answer is, “That is a very complex question. What do you think?” I’ve noticed that among Westerners in the Arab world, this issue can stir up a lot of emotions, which I don’t find puzzling.
Arthur, Eddie. 1993. Speech-Led Versus Comprehension-Led Language Learning. in Notes on Linguistics, vol 60.
Badawi, Elsaid. (2002). The quest for the level 4+ in Arabic: Training level 2-3 learners in independent reading. In Leaver, Betty Lue & Shekhtman, Boris (Eds.), Developing Professional-Level Language Proficiency, pp. 156-176. Cambridge: Cambridge University Press.
Ferguson, Charles E. (1959). Diglossia. Word, 15, 325-340.
 Thank you to David Cross, Becky Huston and Frances Pilavci for their comments and suggestions.
 In terms of all four dimensions, H and L are fundamentally different: cognitively they involve different processes, socioculturally they differ in their early acquisition and later functions, redemptively, they must play different roles, and temporally H has L as a resource already in place when learning of H begins.
 One reader pointed out that many expats who put a heavy emphasis on H from an early point in their Arabic learning do not necessarily become fluent readers of normal H literature, even after a number of years. Perhaps tackling L and H more on their own terms, and at the appropriate points in one’s development, would lead to better outcomes in this regard. It may be that some current approaches to H are cases of too much in the short term and too little in the long term.
 One reader pointed out another practice that basically belongs to the expat subcultures in the Arab world: transcribing Arabic in Roman orthographies. This raises issues unrelated to diglossia, and so I won’t go into them here.