June 23, 2007

Arabic proficiency levels

A few days ago, Spencer Ackerman quoted a Q&A ("U.S. Embassy-Baghdad: Y Kant State Kommunikate", 6/19/2007):

There are about 200 Foreign Service Officers in the U.S. Embassy in Baghdad. How many of them do you figure are fluent in Arabic? The question was posed in today's State Department press briefing, and here's the answer:

Question: How may Arabic speakers with 3/3 levels of proficiency are currently serving at Embassy Baghdad?
Answer: We currently have ten Foreign Service Officers (including the Ambassador) at Embassy Baghdad at or above the 3 reading / 3 speaking level in Arabic. An additional five personnel at Embassy Baghdad have tested at or above the 3 level in speaking. A 3/3 indicates a general professional fluency level.

The editors of Foreign Policy commented on their Passport blog ("The State Department's Arabic problem is worse than you think", 6/21/2007):

This is actually more alarming than it sounds. No wonder U.S. Ambassador Ryan Crocker was raising hell.

A 3/3 level of proficiency is virtually useless for conducting serious business in Arabic. The use of the word "fluency" here is deeply misleading: Someone with a 3/3 would not be able, for instance, to do simultaneous translation of a meeting, and would struggle to translate complicated documents. Anything technical, legal, or politically sensitive would not be something you'd want a 3/3 to handle. For that, you'd need someone closer to a 5 or better yet, a native speaker with a large vocabulary and superior writing skills in two languages. Such people are rare, because the amount of investment and time it takes to reach such rarified heights is more lucratively deployed elsewhere.

[Sally Thomason posted some less negative information back in February ("Another view of Americans & Arabic in the Gulf", 2/19/2007), and I've gotten some additional feedback from a couple of readers, which I'll post separately. And you may also enjoy the jokes in a post from 2004, "Iraqi chicken", though I sincerely hope that they are now out of date. Meanwhile, I believe that the information below on the meaning of proficiency scales, and the problem of Arabic languages and registers, remains relevant, whatever the facts and interpretations about proficiency in the Baghdad embassy.]

Those numbered levels refer to the ILR ("Interagency Language Roundtable") Language Proficiency Skill Levels, which rate proficiency on a scale of 0 to 5 for each of five skills, namely speaking, listening, reading, writing, and translation. The five levels are described as 0 = "no proficiency", 1 = "elementary proficiency", 2 = "limited working proficiency", 3 = "general professional proficiency", 4 = "advanced professional proficiency", and 5 = "functionally native proficiency".

As you can imagine, there are elaborate testing materials and procedures designed to evaluate these skills in a reliable way.

In the range of skills under discussion, Reading 3 is described in detail as:

Able to read within a normal range of speed and with almost complete comprehension a variety of authentic prose material on unfamiliar subjects. Reading ability is not dependent on subject matter knowledge, although it is not expected that the individual can comprehend thoroughly subject matter which is highly dependent on cultural knowledge or which is outside his/her general experience and not accompanied by explanation. Text-types include news stories similar to wire service reports or international news items in major periodicals, routine correspondence, general reports, and technical material in his/her professional field; all of these may include hypothesis, argumentation and supported opinions. Misreading rare. Almost always able to interpret material correctly, relate ideas and "read between the lines," (that is, understand the writers' implicit intents in text of the above types). Can get the gist of more sophisticated texts, but may be unable to detect or understand subtlety and nuance. Rarely has to pause over or reread general vocabulary. However, may experience some difficulty with unusually complex structure and low frequency idioms.

Reading 4 is:

Able to read fluently and accurately all styles and forms of the language pertinent to professional needs. The individual's experience with the written language is extensive enough that he/she is able to relate inferences in the text to real-world knowledge and understand almost all sociolinguistic and cultural references. Able to "read beyond the lines" (that is, to understand the full ramifications of texts as they are situated in the wider cultural, political, or social environment). Able to read and understand the intent of writers' use of nuance and subtlety. The individual can discern relationships among sophisticated written materials in the context of broad experience. Can follow unpredictable turns of thought readily in, for example, editorial, conjectural, and literary texts in any subject matter area directed to the general reader. Can read essentially all materials in his/her special field, including official and professional documents and correspondence. Recognizes all professionally relevant vocabulary known to the educated non-professional native, although may have some difficulty with slang. Can read reasonably legible handwriting without difficulty. Accuracy is often nearly that of a well-educated native reader.

Speaking 3 is:

Able to speak the language with sufficient structural accuracy and vocabulary to participate effectively in most formal and informal conversations in practical, social and professional topics. Nevertheless, the individual's limitations generally restrict the professional contexts of language use to matters of shared knowledge and/or international convention. Discourse is cohesive. The individual uses the language acceptably, but with some noticeable imperfections; yet, errors virtually never interfere with understanding and rarely disturb the native speaker. The individual can effectively combine structure and vocabulary to convey his/her meaning accurately. The individual speaks readily and fills pauses suitably. In face-to-face conversation with natives speaking the standard dialect at a normal rate of speech, comprehension is quite complete. Although cultural references, proverbs and the implications of nuances and idiom may not be fully understood, the individual can easily repair the conversation. Pronunciation may be obviously foreign. Individual sounds are accurate: but stress, intonation and pitch control may be faulty.

Examples: Can typically discuss particular interests and special fields of competence with reasonable ease. Can use the language as part of normal professional duties such as answering objections, clarifying points, justifying decisions, understanding the essence of challenges, stating and defending policy, conducting meetings, delivering briefings, or other extended and elaborate informative monologues. Can reliably elicit information and informed opinion from native speakers. Structural inaccuracy is rarely the major cause of misunderstanding. Use of structural devices is flexible and elaborate. Without searching for words or phrases, the individual uses the language clearly and relatively naturally to elaborate concepts freely and make ideas easily understandable to native speakers. Errors occur in low-frequency and highly complex structures.

It seems a little exaggerated to describe "3 reading/3 speaking" as "virtually useless" -- I'm sure things would be better if more than 5% of U.S. foreign service personnel in Iraq had that level of proficiency in Modern Standard Arabic -- but the Passport blog goes on to point out a much more serious problem:

What's more, I would assume that the proficiency scale refers to Modern Standard Arabic (MSA), which is what most students of Arabic learn and is the language used in most newspapers and for Al Jazeera's broadcasts. The dialect spoken by Iraqis is very different from MSA and from other Arabic dialects.

A rough (but fairly accurate) anology would be to see MSA in the role of Latin in 17th-century Europe as the language of formal discourse. As I understand it, the modern Arabic "colloquials" are as different from MSA and from one another as Italian is from Latin or from Spanish. The fact that American diplomats (and soldiers) aren't taught Iraqi (or Egyptian, or Moroccan, or whatever) is a bigger problem than whether they reach proficiency 3 or proficiency 4 in MSA.

For a discussion of these issues from the point of view of contemporary speakers of different varieties of Arabic, see Mohamed Maamouri, "Language Education and Human Development: Arabic Diglossia and its Impact on the Quality of Education in the Arab Region" (1998). On pages 32-42, there's an enlightening discussion of the "linguistic nature of Arabic diglossia".

The "Arab schoolchildren profiles" that Mohamed gives on pp. 27-28 provide a helpfully concrete picture of the situation. Here's one (note that Fusha is an Arabic term for the literary language, including MSA):

Hela is a sixth-grade primary school student living in Tunis. She spends her summers in Nabeul with her grandmother. Her two best friends there are Hiba and Meriem. Hiba lives in Nabeul all year round and is the same age as Hela. Meriem is a year older and lives in La Marsa during the school year. Hela goes to a private school where she started French and Arabic at the same time. She has more than 20 hours of classes in Arabic and about 10 hours in French a week. All the subjects other than French, such as Math and Biology, are taught in Fusha. Sometimes the teacher explains things in Arbi, but the students often have to speak in Fusha. Hela does not like Fusha as much as Arbi, it feels too alien to her. She even likes French better than Fusha. Meriem’s classes are a lot like Hela’s. She prefers French and often uses French words when she’s speaking Arbi. She thinks it makes her sound cool, like an adult. Hiba, on the other hand, didn’t start French until the third grade. Even though she now has the same number of hours of each language as Hela does, she prefers Arabic (both fusha and Arbi) to French and reads more Arabic books.

The three girls play together and watch television. Their favorite shows are Saoussen, which is in Fusha, and Les Schtroumfs, which is in French. Sometimes, when they play, they pretend to be the cartoon characters and try to sound like them. Hiba likes playing Saoussen best, because she doesn’t play well when they speak in French. Meriem prefers Les Schtroumfs because her Fusha is poor. They usually just speak Arbi together. After the summers over, Hela and Meriem go back to their homes. They decide to write each other letters over the school year. After the first day of school, Hela runs home to write letters to her friends. She starts to write a letter to Hiba, in Fusha, but feels that this is not a friendly letter. It feels more like homework. She thinks in Arbi, but cannot write what she means, and has to translate. Frustrated, she decides to write to Meriem first. She quickly realizes that her best bet is to write in French, but still struggles with finding the right words to say what she means. Finally, she settles on using Arbi words that she approximates phonetically and finishes one letter. For Hibas letter, though, its harder for her to do this with Fusha, so she just writes a very short letter and writes some words in French. These solutions work, but leave her feeling unsatisfied. She feels closer to Meriem because she can communicate with her better. She rapidly loses interest in writing to Hiba, though.

Hela's cousin, Farah, grew up in Saudi Arabia. She is the same age as Hela and is in the fourth grade. Farah only speaks Saudi Arabic, Fusha, and English, which she studies at school. She feels that Fusha is strange and silly. Nobody really speaks it there either. When Farah and Hela get together, they can only speak a mixture of their dialect with Fusha. It is very strange for both of them. They hardly ever write each other letters, because they’d have to do it in Fusha, which neither feels comfortable with. Farah feels resentment towards Fusha and reads even less. She doesn’t like music in Arabic as much as English or French music and only reads in Arabic if it is mandatory. Her French continues to improve and her Fusha remains poor. This does not bother her though, because she knows that once she gets to secondary school, Fusha would be much less important and if she wants to be a doctor when she grows up, she will only need French.

And here's an anecdote from p. 33:

Parkinson relates the story a friend who was a passionate supporter of fusha and who decided to stick to it exclusively in his family in order to give his children the full advantage of having it as a native language. Getting on a busy Cairo bus with this friend and his three-year-old daughter, the two of them, father and daughter, were separated and the yelling that was necessary to reestablish the contact took place in fusha making the entire bus burst out in laughter.

On the other hand (p. 38):

The superiority that Arabs bestow on their heritage language leads to a quasi-general denial of the existence of a home language, in this case colloquial Arabic. Arabs consider in fact that what is spoken at home, and elsewhere in common daily activities, is merely incorrect language which is only acceptable because it deals with lowly functions and topics. There is a prevailing feeling among Arabs that their language is imbued with a natural superiority. This ‘prestige valuation’ of fusha is explained by Arabs as relating to such qualities as beauty, logic, and a high degree of expressiveness. Fusha carries in its own etymology the myth about its eloquence and high degree of correctness. Moreover, Arabs despise the spoken colloquial forms and even deny that they use them because they consider the colloquials they speak as ‘degraded’ and corrupt forms of the language. They give them derogatory names such as barbri “barbarian” or yitkallam bi-l-fallaaqi “he speaks the language of woodloggers.”

This situation makes the task of foreign learners more difficult, since they need to learn to deal appropriately with a very broad range of mixtures of "high" and "low" languages. This is true to some extent in any language, but the range of diglossia in "Arabic" appears to be significantly greater than in most other modern situations. You need to imagine a situation in which "Latin" is used to refer not only to classical and patristic Latin, but also to the spoken versions French, Italian, Spanish and Portuguese (with none of them having any standard written form).

The sociolinguistic situation in Iraq is different in many ways from the situation in North Africa, but it doubtless remains true that Fusha "feels alien" to most Iraqis, even if they are able to understand it and to speak it to some extent. At the same time, formal settings require formally correct language.

[Update -- Joel Thibault writes:

It may interest you to know that Les Shtroumpfs is well known to English speakers as The Smurfs.


[Update #2 -- Lameen Souag writes:

Great post. A minor point though, relating to Mohamed Maamouri's article: "barbri" (bəṛbṛiyya, it would be in Algeria) is not so much "barbarian" as to "Berber" - as far as I know, it's only used to refer to North African Arabic (although it's nearly obsolete in that sense in my area), and hence to dialects of Arabic adopted by what were ethnically largely Berber populations. In Moroccan dialects (I think), and commonly in Fusha (eg in ibn Khaldun), it means "Berber". It still has a derogatory origin, of course - the ethnonym "Berber" does probably derive from "barbarian" - but this particular usage probably derives from the ethnic sense, rather than directly from the derogatory one.


[Update #3 -- Nathan Wagner writes:

While I don't know what the current situation is, I attended DLI (the Defense Language Institute, which is where the military trains its foreign language speakers) in 1988-1990 where I took the basic Arabic course. In contrast to your claim, everyone took a 47 week basic course in MSA, followed by a 16 week dialect course. As I recall, the dialects then were Syrian and Egyptian (though I was assigned Egyptian and that is the only one I am certain about). While I hope that the dialects have changed since then, I would imagine that the basic curriculum is more or less the same. A brief look at their website isn't very informative, but they do note a 63 week basic course in Arabic, which would correspond to what I took. Interestingly however, there is no actual mention of specific dialect training.

I also don't know what the current situation is. However, as of a few years years ago, U.S. military personnel being trained at DLI for duty in Iraq apparently did not learn any colloquial Iraqi, and as far as I know, no significant amount of any other colloquial Arabic either. As I wrote in "Iraqi chicken" (7/15/2004),

Among the interesting points that yesterday's speaker (a captain in the Army reserves) made:

Her DLI (Defense Language Institute) training in Arabic was less useful than she would have wished, mainly because it was in MSA (Modern Standard Arabic), whose relationship to Iraqi Arabic is roughly like the relationship between Latin and Italian. As a result, as she put it, "they could understand me but I couldn't understand them". I've heard that DLI used to teach a number of modern Arabic languages (often called "colloquials" or "dialects"), but stopped some time ago because the military's personnel system couldn't deal with the distinctions. As far as the personnel system was concerned, Arabic is Arabic; but sending someone trained in Moroccan Arabic to (say) Kuwait is like sending a Portuguese speaker to Romania.So DLI decided to stick with MSA, which is the language of formal discourse throughout the Arab world, though it's no one's native language. I don't know whether this is the reason, but it's certainly true right now that DLI teaches MSA rather than the local languages.

In general, I believe that DLI does an excellent job, and the people who run the place and determine the curriculum are certainly aware of all these issues, so some changes have probably been made over the last few years. Without any specific knowledge of what's happened, though, I would guess that providing good course materials and recruiting enough teachers fluent in Iraqi (since few of their existing Arabic instructors were from Iraq) would both have been challenging problems. ]

[We should also note that there are several different varieties of colloquial Arabic spoken within Iraq -- Ethnologue distinguishes Mesopotamian Spoken Arabic (estimate 11.5 million speakers in Iraq, 15.1M in all countries), North Mesopotamian Spoken Arabic (estimate 5.4M speakers in Iraq, 6.3M in all countries), Najdi Spoken Arabic (estimate 900K speakers in Iraq), and Gulf Spoken Arabic (estimate 40K speakers in Iraq). That doesn't count the 3.3M speakers of various kinds of Kurdish, and smaller numbers of speakers of more than a dozen other languages.]

Posted by Mark Liberman at June 23, 2007 10:31 AM