The Nature of Old Arabic and its Change into Middle and then Modern Arabic

Table of Contents

I.1 Introduction To Major Issues

In this paper we will examine the various theories about the development of Arabic. This includes the nature of the language prior to the onset of Islam, what happened to it as it spread out from the Arabian Peninsula and what was the nature of the changes that occurred to Arabic once it became the dominant spoken language in the Middle East and North Africa. Finally, how did the Modern Arabic dialects develop from the Arabic that established itself after the Islamic expansion? Further questions arise depending upon the answers to the preceding questions. Was the Arabic in the Arabian Peninsula in the early 7th century for all intents and purposes a unified whole or was it many dialects with a poetic form alien to any existing spoken form of the language? Did the current dialects develop as the result of the conquered subjects in the new Islamic empire being unable to learn the common poetic language or did the conquered populations learn the different spoken dialects of the conquering armies? Posing the preceding question with a slightly different emphasis, had the break between the spoken language and the written language already occurred? On another tack, did the process of Arabic suddenly being spoken by large numbers of non-natives speakers, who did outnumber the native speakers of Arabic, produce a drastic change in Arabic regardless of the relationship between the spoken the dialects and the poetic language? Return to Table of Contents

I.2 Terms Defined

This paper will summarize the views of prominent scholars on this question and suggest possible areas of further inquiry, speculation and debate. For purposes of this paper I want to define a few terms in the following way: To review, the types of known and postulated forms of Arabic discussed in this paper are: PK, OSA, Al-cArabiyya, WMA, SMA, NSA, MSA, KII and KIII. As anyone who has tried to learn Arabic knows, defining Modern Arabic is not an easy task. There are at least 4 major dialect groups of varying degrees of mutual intelligibility. Each dialect differs from the written form to such a degree that a person from the US who only learns Modern Standard Arabic will arrive in the Arabic-speaking world with virtually no useful conversational skills. While most Arabophones would probably understand such a person just fine, he/she would be lucky to understand one reply in twenty. However, MSA is perceived to be the real language, and no one will be insulted by a skilled attempt to speak MSA. A surface explanation for this is that Al-cArabiyya became locked in as the language of the Quran and great effort went into maintaining it as the linguistic ideal, which amounted to forging it in stone. The dialects on the other hand had no official status and could therefore develop without restraint. This combined with ever decreasing literacy rates and a decline in intellectual and mercantile activity in the period prior to and coincident with the European expansion of the 18th and 19th century led to a general atomization of the Arabic dialects. It seems as if the one thing upon which all informed sources agree is that this explanation is inadequate to explain in-depth the situation for Arabic.Return to Table of Contents

II Linguistic Situation in Pre-Islamic Middle East

I cannot help but throw in a few words of caution here. The situation for Arabic as it spread out from the Arabian Peninsula at the time of the Islamic conquests was quite complex. In addition the linguistic situation of the conquered territories was if anything even more complex, especially in the core area of the Middle East, which is Egypt, the Levant and Mesopotamia.

In Egypt the common language was some form of Coptic, which was a direct descendant of the Egyptian language used by the Pharaohs. Coptic was also used as a liturgical language by the Egyptian Christian Orthodox Church. Egypt was also part of the Eastern Roman Empire or the Byzantine Empire as it is variously called. The Hellenistic culture in Egypt went back 900 years to when Alexander the Great conquered Egypt. What this means is that not only was Greek used as the language of administration there was also a Greek speaking population which lived in Egypt at the time of the Islamic conquest.

The situation in the Levant was even more complex. The Jewish people spoke Aramaic as their native tongue, but used Hebrew as their liturgical language. Some of the Christians had developed Syriac, a special form of Aramaic which served as both their literary and liturgical language, but spoke some dialect or other of common Aramaic. Greek was the language of the administration. In the areas which bordered on the desert in what is now the Sinai Desert of Egypt, the Negev Desert of Israel and most of Jordan and the Golan region of Syria, the Ghassanids spoke Arabic.

In Mesopotamia Pahlavi was spoken by the population and was also the official language of administration. Pahlavi is the ancestor of modern Persian. Hebrew was the liturgical language of the Jews, and Aramaic was spoken by some elements of the population. Syriac was used as a liturgical language by the Christians and Arabic was spoken by the Lakhmids. I think that it is very interesting that in both the Byzantine and the Sassanid empires in Iraq/Persia there were Arabs gradually filtering in from the Arabian Peninsula.Return to Table of Contents

III Theories about Pre-Islamic Arabic

Many differing conjectures have emerged about the linguistic situation at the time of the prophet and the inception of Islam. Each explanation attempts to describe the linguistic situation in the Arabian Peninsula just prior to the Islamic expansion.Return to Table of Contents

III.1 Summary of Chaim Rabin

Chaim Rabin puts forth the view that at the time of the Quran there was already a distinct linguistic split between what he calls the Poetic KoinŽ (PK) and the different tribal dialects (OSA). The Poetic KoinŽ was not a spoken language at all, but had evolved strictly as a vehicle for the poetry at the fairs. Alongside the PK there were different tribal dialects of Old Spoken Arabic with varying degrees of mutual intelligibility. This Poetic KoinŽ which Rabin postulates was already a literary language, was used by Meccans for business, and by proselytizing Christians. Nowhere does Rabin argue that the diglossia situation was so bad that the Poetic KoinŽ had to be treated as a separate language from the tribal speech systems.Return to Table of Contents

III.2 Summary of Corriente

Corriente's article focuses primarily on the evidence of the early Arabic grammarians, and his major goal is to examine the pile of evidence in an attempt to deduce the source of the Poetic KoinŽ. He states, almost as an aside that the urban dialects of Syria-Palestine and Iraq might have been derived from the i'rabless Nabati Arabic, which had crept into Syria and Iraq prior to the Islamic expansion. Based on this, the question to ask is how far north and west had Nabati or Lakhmid/Ghassanid Arabic spread before the advent of Islam? Corriente uses the evidence of the grammarians to conclude that the declensional endings were still in use in the Bedouin dialects for two hundred years or so after the Islamic expansion from the Arabian Peninsula. He concludes that there were two main dialect groups, Eastern and Western, both of which were different from the Poetic KoinŽ. Corriente's assertions and arguments most closely resemble the general view expressed in the lectures in class by Professor Cadora. Corriente makes a pretty strong case that Old Arabic (i.e. pre-Islamic Old Spoken Arabic + the Poetic KoinŽ) was in flux and was not as uniform as the Classical Arabic that became codified by the grammarians during the early Abbasid era. In fact he posits the existence of i'raab bearing and i'raabless Arabic existing at the same time, if not in the same place. His whole article is built around the premise that Classical Arabic did not become codified until the 8th or 9th century.Return to Table of Contents

III.3 Summary of Versteegh

Versteegh argues that the Old Spoken Arabic and the Poetic KoinŽ which existed at the time of the prophet and shortly thereafter was a single language which he calls Old Arabic. Any variation which existed was not beyond the range of the normal linguistic variation found in any living language. His basic position is that the descriptions and conclusions of the early Arab grammarians about the Arabic language are reliable and should be used as real evidence for determining the state of early Classical Arabic. His strongest argument for the unity of Old Arabic is the evidence of the early grammarians. In his first chapter he does not state how quickly the changes take place, but he does characterize these changes as radical and far-reaching. Versteegh asserts that the changes between Old Arabic and New Arabic (his terminology) are so deep that New Arabic constitutes a new language type. Blau, as well, argues that Old Spoken Arabic and the Poetic KoinŽ were in essence the same language. His evidence is the lack of pseudo-corrections in the Quran. He also argues that the poets would not have been able to correctly compose the Jahiliyya poetry if they were not comfortable using the features of the Poetic KoinŽ. What Versteegh and Blau argue for, using scientific terminology, is very reminiscent of the received truth about Arabic in the Arab world. The pure Arab mother tongue got altered irretrievably by the muwallad who could not master its complexities and subtleties.Return to Table of Contents

III.4 Summary of Zwettler

Zwettler takes the position that the Poetic KoinŽ was a very ritualized form. In this regard it could be compared with the Greek in which Homer composed the Iliad and the Odyssey. The features of this kind of poetic language used in a tradition of oral poetry are 1) the preservation of archaic features, 2) the formulaic use of set phrases and 3) the borrowing of virtually synonymous words from other dialects in order to preserve rhyme and meter. Viewed in this way one can only conclude that the Poetic KoinŽ was a hodgepodge containing many features which never coexisted at any one time, in any one dialect. It would thus be full of archaic forms from every era and haphazard borrowings from every dialect and could never represent any single stage of any single dialect.

Zwettler goes on to argue that even though the poets could compose poetry correctly using i'raab, this does not prove that they spoke the Poetic KoinŽ as their native language. To drive home this point he also presents the argument that the later poetry tended to use, more often than in the old poetry, rhyming schemes that did not depend on knowing the i'raab well. The point here is that the later poets didn't necessarily have such a good knowledge of the grammar of language of the Poetic KoinŽ. Only a very specialized expert composed poetry, and the formulaic nature of the poetry assured the poet of being able to mimic older formulas which came equipped with the correct i'raab. He further casts doubt on the validity of the grammarians' accounts about the "purity" of Bedouin speech by contending that the grammarians only sought out those Bedouin informants who knew how to recite the oral poetry. What we actually have is not representative samples of natural Bedouin speech, but rather samples from Bedouin informants trained in producing the old oral poetry.Return to Table of Contents

III.5 Summary of Ziadeh

Ziadeh uses the evidence of the multiple forms of broken plurals to argue that extensive cross dialect borrowing occurred in Jahiliyya poetry. He also argues that sometimes the poet would just invent forms in order to get the correct rhyme and meter. He seems to support Zwettler's argument that the Poetic KoinŽ was a language that became increasingly artificial during the pre-Islamic period. Zwettler and Ziadeh both seem to be aiming their arguments at Blau, in an effort to show that the use of "correct" Poetic KoinŽ does not mean that the Poetic KoinŽ was the poets' native vernacular.Return to Table of Contents

III.6 Summary of Cadora

Cadora makes the point that in the history of the "culture area" there is a persistent and consistent process whereby groups of people change ecological structure. Linguistic change always accompanies these changes in ecological structure. The three structures in the Middle East are Bedouin, rural and urban. The general trend is for a group to move from Bedouin to rural and then from rural to urban, with intervening transitional structures in between the three main structures. For instance a band of Bedouin herders may chance upon a particularly choice patch of grazing land with a permanent water supply, and since there is no reason to move on, they have by doing this opted to become sedentary herders. It can be argued that this transitional stage must occur when making the transition from nomadic herder to settled farmer. The next phase after having become settled herders is to plant some beans, watermelons and/or wheat. The pattern can run in reverse as well, i.e. a group can go from urban to rural and eventually end up as Bedouin, perhaps as the result of a major catastrophe. However, the overall trend is to move from Bedouin to rural and then from rural to urban. The point is that the linguistic variation in pre-Islamic Arabia was at least as dependent on the ecological structure of the speech community as it was upon geographical location or tribal affiliation of the speech community. Linguistic change, both external and internal, always accompanies a change in the ecological structure of the community. In Cadora's view there was an inherent variability of the Arabic language based on community type. This situation has existed for as long as there has been anything that can be called Arabic.Return to Table of Contents

III.7 Table of Positions on Critical Issues

Return to Table of Contents

IV Nature of Post-Islamic Changes in Arabic

This leads us directly into debate over the nature of the changes which occurred during and after the Islamic conquests. How did Arabic change after the conquests? What, if anything happened to OSA and the Poetic KoinŽ as they developed into Al-cArabiyya and Spoken Middle Arabic? How does the history of these developments affect the modern linguistic situation of the Arabophone world? What is the developmental relationship between New Spoken Arabic/Modern Standard Arabic and OSA/PK?Return to Table of Contents

IV.1 Diglossia Always Existed in Arabic

On the one extreme we have the scholars who claim that PK and OSA have always been different from each other and the PK is a created form. These scholars also claim that the modern dialects (NSA) exhibit features which are a continuation of tendencies which existed in OSA. In this general view diglossia in Arabic pre-dates Islam. The scholars who want to show that diglossia pre-dates Islam want to show that the declensional endings for mood and case (i'raab) had already been lost at the time of the prophet Mohammed. This loss of the declensional endings is sometimes also described as spoken Arabic's change from being a synthetic language, to being an analytic language. In this regard neither Zwettler nor Ziadeh put any effort into explaining the diglossia situation in Modern Arabic (NSA+MSA) other than to suggest that it existed all along and therefore needs no further explanation.

Corriente and Rabin both put great effort into trying to prove that the i'raab had been lost in common speech by the early 7th century CE. They do not directly address the issue of what other changes occurred concurrently with the loss of the i'raab. They are obviously of the opinion that everything else follows from the loss of the i'raab. If pre-Islamic OSA did not have i'raab then diglossia always existed, and the great gap between the spoken and written forms today is strictly due to the arrested development of MSA and the continued evolution of the dialects which has spanned a period of 1,416 years.Return to Table of Contents

IV.2 Old Arabic Was Essentially One Language

Versteegh and Blau very definitely state that OSA and PK were one and the same language. Versteegh hedges a little bit and talks about registers and special forms for special uses but he clearly states his belief that the these two forms were complementary to each other and cut from the same cloth. Versteegh's theory is pretty involved but briefly summarized states that the varieties of New Arabic are all based on innovations and processes which occurred after the Islamic conquests. Blau doesn't go into so much detail but insists that we must not overestimate any alleged difference between the Poetic KoinŽ and the tribal vernaculars. Blau, like Versteegh, is very clear that the changes occurring between Old Arabic and Middle Arabic, to use Blau's terminology, are far reaching and occurred as the result of the Islamic conquests. They both say that Old Arabic and the post-Islamic Arabic are two different kinds of languages. Blau's article "The Beginnings of the Arabic Diglossia. A Study of the Origins of Neoarabic" seems to be mainly aimed at going after Ferguson's proposed KoinŽ II theory.Return to Table of Contents

IV.3 Ferguson's Theories of Diglossia and Koiné II

Ferguson in his article "Diglossia" is attempting to define diglossia in general, and does not limit the discussion to Arabic alone. His thesis is that diglossia generally occurs in situations where 1) there is a large body of literature to which the community is very attached because it is culturally defining, 2) literacy rates are low. and 3) the literature has been around for a number of centuries. He also posits that diglossia tends to be relatively stable. He wrote another article called "The Arabic KoinŽ," which both Versteegh and Blau critiqued sharply. In this article he argues that diglossia was well developed at the time of the Islamic conquests, but that the conquests caused a linguistic leveling or new KoinŽ (KoinŽ II). KoinŽ II was used by the military for inter-tribal communication and for communicating with non-Arabs in the military camps in the newly conquered territories shortly after the Islamic conquests. The modern dialects which developed outside the Arabian Peninsula are all descendants from this new KoinŽ II rather than being derived from Al-cArabiyya or the Poetic KoinŽ. He makes a pretty strong case that the dialects outside the Arabian Peninsula must have had a single source.Return to Table of Contents

IV.4 Koiné III?

It was also conjectured during Cadora's Winter '96 lectures at the University of Michigan that there may have been a verbal pre-Islamic KoinŽ III. This Arabic was heavily influenced by Nabatean Arabic and was used for inter-tribal spoken communication on caravans of a mixed tribal nature and in major commercial centers, especially to communicate with non-Arabs or Arabs who were already living outside of the Arabian peninsula. This Arabic had lost the i'raab and may even have been a pidginized form. It can be conjectured that it played a role in the formation of Ferguson's KoinŽ II.Return to Table of Contents

V Analysis of the Theories

In my opinion there is a lot that needs to be explained, no matter which theory we choose. So far no particular theory explains everything to my satisfaction. I agree with Versteegh when he says that the disappearance of the declensional endings is the least of the changes. As Rabin points out, one cannot convert dialectal Arabic into Fusha by appending the declensional endings.Return to Table of Contents

V.1 Lack of Hard Evidence

Several hurdles immediately present themselves when we attempt to gather information about the state of pre-Islamic Arabic. To begin with, the Arabic script was borrowed from Nabatean. The Nabateans had originally used it for writing Aramaic, not Arabic. This script did not have all the of the consonants that existed in Arabic. The dots which differentiate between t,b and th or j,H and x had not been standardized yet. Thus the script is ambiguous. Another point about the script is that it was written without marks for showing the short vowels. As if all of this wasn't bad enough, the Arabs of the peninsula tended to write on perishable items, such as palm fronds or tree bark. There are very few pre-Islamic Arabic writing samples. To make matters worse almost all of the Arabian Peninsula has been closed to archeological investigation until very recently. The vast majority of the samples available carry almost no information about one of the most hotly debated issues, namely was there i'raab or not? Blau analyzed a sample of Arabic that had been fortuitously written with Greek script from the 8th century CE in which the i'raab were absent. Even in this case this is not certain evidence for the loss of inflectional endings. The scribes using the Greek letters may have been merely mirroring the practice in Arabic of not marking the endings. So we have disputable evidence that if the i'raab were present prior to Islam, they may have disappeared very shortly afterwards.

The point here is that there is very little hard direct evidence. We are forced to reconstruct and make conjectures and see what fits into this puzzle which has many missing pieces. In the final analysis we must admit that we may never know with any certainty the truth about how Modern Arabic was formed from the earlier stages of the language.Return to Table of Contents

V.2 The Evidence That We Do Have

Having said all of that let us examine the arguments in the light of what evidence there is. My strongest argument against the inherent unity of old Arabic is that the all of the urban dialects outside the Arabian Peninsula have a common way to negate verbs and nominal sentences. The verb is negated by surrounding the verb with ma .... sh. To negate a nominal sentence one prefaces it with some variant of mish or mush. It is true that in the peninsula other rules apply. For this reason I like Ferguson's article and chart #10 that was handed out in class with the Poetic KoinŽ and KoinŽ II. Classical Arabic formed in parallel with KoinŽ II sometime around the formation of Islam and the modern dialects are formed out of KoinŽ II. It explains how a change in such a fundamental feature as verb negation could have happened to all non-peninsular dialects, even though they all have had separate evolution. Versteegh's position that there was only one language at the time of the Quran does not adequately account for this feature. Blau's explanation that this development is recent and caused by diffusion does not adequately account for why this feature is in nearly every non-peninsular urban dialect.

I do think that Versteegh's arguments in chapter 2 for the abruptness of the transition from Old Arabic to New Arabic are very persuasive. The dialects exhibit fundamental structural differences from the Fusha. If we are to believe the Arab Grammarians from the 9th century, the common urban language was already hopelessly removed from the standard of the Quran and the pre-Islamic poetry. How different from what Corriente and Rabin are saying is Versteegh's assertion that there was only one language at the time of the prophet? The question that needs to be answered is, were the dialects mutually intelligible with each other and the Poetic KoinŽ? If yes, then I think I can allow Versteegh his assertion that there was only one language with different "registers" as he calls them. Even without the mutual intelligibility of the pre-Islamic dialects there still seems to be a real concern on the part of the grammarians that if they didn't do something quick that the meaning of the Quran would be lost. This would seem to indicate that the language had indeed undergone a noticeable and therefore radical change in a very short period of time, merely 200 years. On this point of mutual intelligibility, I did observe when I was in the Middle East that Egyptians and Jordanians could talk to each other in dialect, and as far as I could tell they weren't adjusting their dialects for each other. This points to another interesting consideration not discussed in any of the literature reviewed here, namely that the Arabic language community seems capable of handling a certain amount of inherent variability. The problem exists mainly for me as an outsider. Due to the fact that I haven't been exposed to this variability since birth, I have a lot of catching up to do when thrust into this environment.

I think there are many issues that are still debatable. If Arabic did not undergo rapid change, why then are some of the features of the dialects which differ from Old Arabic identical with each other? If New Arabic did evolve slowly from a pre-existing situation of diglossia, why then were the 9th century grammarians so intent on "preserving" the language of the Quran? Corriente mentioned them in passing, but what was the role of the Nabateans if any, in the spread and development of New Arabic?

So how do we make sense out of all this? Obviously some very knowledgeable people have devoted their lives to arguing these issues. For one thing everybody seems to be arguing as if each theory is mutually exclusive of all other theories. I want to point out that Versteegh's theory of pidginization and creolization followed by gradual decreolization does not necessarily rule out Ferguson's KoinŽ II. The presence or absence of i'raab in Old Spoken Arabic does not rule out the existence of a poetic KoinŽ that was understood, in a passive way by all, but produced only by the tribal rawi.Return to Table of Contents

V.3 A Proposed Scenario

Based on my current understanding of the evidence summarized so far in this paper, I think that the following scenario seems to fit the facts as I know them from the class readings and lectures. There was diglossia at the time of the prophet, or more precisely, there was a high degree of linguistic variability in Old Arabic. Probably a noticeably higher degree than found in most languages but not as severe as it exists in New Arabic. Specifically the different dialects were less diverse than they are today, but having said that there was no dominant dialect nor linguistic center. Mecca may have been in the process of becoming the dominant cultural linguistic center but it had not yet achieved full dominance. The linguistic variability was defined in terms of ecological structure and to a lesser degree by geography and then by tribal affiliation. In any event we cannot use Ferguson's definition of diglossia, unless we also define the Jahiliyya poetry as being a defining set of literature. The existence of a low literacy rate is the second point of Ferguson's diglossia definition and it doesn't quite seem to apply. Literacy rates were low, but the exact place of writing in pre-Islamic Arabia is not clear.

The Poetic KoinŽ was not representative of any single dialect or any single era. It had higher status than any dialect. It was not yet as standardized as Al-cArabiyya was to become. It was used for the qasidas, oration and fortune telling. It had existed for at least two centuries if not much longer. The Quran was an innovative use of this form of the language. There may or may not have been a "market language" (KoinŽ III). There was evidence of writing but the only samples that have survived are grave markers and other inscriptions on monuments. If we could obtain samples of any possible written commercial records it would undoubtedly reveal to us a lot about 1) the existence of a market language and 2) the syntax of the lower status languages in pre-Islamic Arabia.

The Islamic conquests put the Arab armies in charge of a very large empire. Arabophones were vastly outnumbered. The new converts to Islam learned Arabic in very informal and untutored settings. According to both Goldschmidt in A Concise History of the Middle East and Hourani with A History of the Arab Peoples, Arabic did not replace the native languages in the conquered territories for at least 200 years. What this means is that the number of non-native speakers of Arabic vastly outnumbered the native ones. There was also the phenomenon under the Umayyids of the military town kept separate from the local population. This combined with the wholesale taking of local wives produced whole generations of children who were nominally Muslim and Arab but who learned their first language from their mothers, who most likely didn't speak Arabic very well. Arabic was also being used for writing laws, explaining theology, and translating Greek science and philosophy. New words had to be coined or borrowed. This is not the ordinary course of ecological change. All of the Arabs whether Bedouin, farmer, or urbanite had become almost overnight the ruling class of a huge empire. Why shouldn't linguistic change accompany this momentous change in ecological structure? To claim that this upheaval in the social and political fabric of the Arab peoples did not make a mark on their language seems absurd.Return to Table of Contents

V.4 Versteegh vs. Ferguson

I want to combine Versteegh's explanation with Ferguson's explanation. The radical changes which occurred to Arabic took the form first of abrupt pidginization and creolization which was followed by a long period of gradual decreolization. The spoken language to which this process of pidginization, creolization and then gradual decreolization occurred was Ferguson's KoinŽ II. I think this accounts for the features that are common to all dialects but differ from Old Arabic, which is what Ferguson was aiming at. It also accounts for those features where each dialect is different from Old Arabic and is different from the other dialects as well, which is what Versteegh was aiming at. In fact Versteegh uses this anomaly to try to discredit Ferguson's theory.

I differ with Versteegh's analysis where he wants to say that the dialects developed out of Old Arabic, and that this Old Arabic was the same language as the Quran. Ferguson's arguments in his article "Grammatical Agreement in Classical Arabic and the Modern Dialects" are very convincing.

Ferguson points out that the only category for which duals exist in modern Arabic dialects is the noun, and that it invariably takes plural agreement. This is very different from Old Arabic and MSA which have dual categories in the verb, pronoun and adjective. In MSA and Old Arabic a dual noun takes dual agreement with the verb, pronoun and adjective which makes it a separate category from singular and plural. In MSA there is a complex set of rules for verbal agreement, and adjectival agreement with the noun. If the verb precedes the noun it takes gender but not number agreement. Non-human plurals take feminine agreement with adjectives. Dual nouns apparently break this rule by invariably taking dual agreement with adjectives regardless of their human/non-human status, but verbs which precede dual nouns only take gender agreement. In the New Spoken Arabic dialects adjectival agreement with non-human plurals can be plural or feminine. It is generally safe in the dialects to give feminine adjectival agreement to non-human plural nouns because even if this is not correct for that dialect it can be taken as a "classicism". However, in the modern dialects dual nouns always take plural agreement. It is hard to imagine that this exact feature could have developed independently in so many different places. Ferguson's argument is very persuasive that the New Spoken Arabic dialects have a common source and that this source is different from the Poetic KoinŽ.Return to Table of Contents

V.5 Ann Miller

Ann Miller in her article "The Origin of the Modern Arabic Sedentary Dialects" in Volume 19 of Al-cArabiyya, 1986 also makes the suggestion that we should combine theories. Although I don't agree with all of her conclusions it is very interesting that other people surveying this topic have proposed combining selected theories. She proposes combining Corriente's theory with those of Blau and Cohen. Her basic position is that New Arabic pre-dates Islam and that there is no solid evidence for a common origin of the modern dialects. However she also states that we may never know the truth about the origin of the Arabic dialects, until more evidence surfaces. A major part of her article is devoted to compiling a list of the arguments against Ferguson's KoinŽ II. Without double checking her sources I will reiterate that I find Ferguson's arguments very persuasive about the common source for the non-peninsular sedentary dialects.Return to Table of Contents

V.6 The Nabateans

It seems likely that the Nabatean style of Arabic had more than its share of influence on KoinŽ II. The first areas conquered were inhabited by Arabs who spoke Nabatean Arabic. The Arabs had controlled the desert trade routes out of Arabia into Damascus and Jerusalem. These places must have had Arab speaking populations as well, who spoke this Nabatean variety. I think these people must have occupied a special place, because they were more or less linguistically competent in both communities. I don't know what the evidence is with regards to their participation in the Islamic conquests after they had themselves been conquered. I still think that as potential linguistic intermediaries between the conquering Arabs and the other conquered peoples they were in a unique position to leave their mark on the form of the language that the conquered peoples learned. Furthermore, I want to point out that they were the source of the Arabic script and they had at one time controlled the trade out of the Arabian Peninsula into the Mediterranean. Their influence on shaping pre-Islamic Arab culture has been underestimated in my opinion. How to prove it is the question here.Return to Table of Contents

V.7 Is I'raab Important?

In all of the preceding discussion people have argued back and forth about the importance of the loss of i'raab and the importance of when it happened. This is the litmus test for when Arabic started to change and how much influence the conquests had on the development of the dialects and the atrophying of Al-cArabiyya. This is also the dividing line between Old Arabic, the synthetic language, and New Arabic, the analytic language. Allow me to point out that Arabic still has many synthetic features. It may not have case endings or verbal mood endings, but it still does have bound morphemes. If anything some items that were not clitics in Al-cArabiyya are now bound to the modified lexical item in New Arabic. For instance in every dialect I know the indirect object 'li' is now an enclitic which is attached to the verb, since it affects stress and intervenes betweeen the negating 'sh' morpheme. To illustrate, "he wrote to me" is "katabli" with the accent on "tab". "He didn't write to me" is "makatabliish", with the accent on the "liish" part of the form. Anyway it seems likely that some of the urban varieties with close contact to the Levant had already lost the i'raab without really leaving the linguistic community of the Poetic KoinŽ. The ambiguity of the script makes it seem to me that the i'raab were part of the variation. That is to say pausal forms were read in connected speech by certain parts of the community. If everybody read texts fully vowelled why was the 'n' of indefinite tanwin not included in the script?Return to Table of Contents

VI Conclusions and Further Questions

Here in twentieth century North America we live in a relatively linguistically homogenous society. Diglossia, and linguistic variability seems scary to us. In the Middle East this is nothing new. The territories conquered by Arabs, as I have mentioned previously, were not linguistically homogenous. They all had at least two important languages other than Arabic in use at the time of the conquests. Furthermore throughout history, the written forms were the domain of a very few specialists, priests and scribes. That the written form was different from the spoken form, was not very important. The modern ideal of universal literacy and the accompanying idea of a unified national language which is closely bound to the written form, is a recent innovation, which developed in Europe during the reformation. Generally the invention of the printing press is given credit for this innovation, but I also think that Protestantism's insistence on personal knowledge of the faith may have had a hand in developing the vernaculars into written languages as well. In a certain sense Arabic was, until very recently, operating according to the old linguistic paradigm. The literary language was the domain of the educated elite, thus, it didn't matter that it did not reflect the linguistic reality of the vernacular forms.Return to Table of Contents

VI.1 Is Diglossia the Correct Term for Arabic?

Let us re-examine the concept 'diglossia' with respect to Arabic and see if it is an appropriate label for Arabic's linguistic situation. Diglossia as defined by Ferguson is the word for the linguistic situation where two different linguistic varieties live side by side. One form is the "high" variety and the other is the "low" variety. The model for this situation is that there are two distinct but related languages which operate in the same linguistic community. One variety is the "native" language and the other variety is an acquired variety which is learned in school. The learned variety has more cultural status than the "native" variety. In this model of diglossia, the speaker is in essence bilingual and must possess two complete grammars in his/her head in order to function linguistically in that community.

Does this situation accurately describe the reality of Arabic? I have yet to talk to the Arab who views his/her dialect and MSA as separate languages. If we build our model taking into account this personal perception, what we get is that each individual speaker has only one grammar. This grammar has lots of sociolinguistic production and parsing rules which account for the inherent variability of the language. One person's personal grammar may be very different from another person's personal grammar with education level being as important a factor as geographic location, ecological niche or religious affiliation.

Following this line of reasoning then, we might find groupings of people with similar personal sets of grammar rules. We could create categories of people whose rules for the "low" variety are very similar to each other. These categories would tend to follow geographic boundaries and to some extent ecolinguistic categories. In contrast, we could separate out people into groups of those whose rules for the "high" variety are nearly identical. These groups would tend to transcend geographic boundaries and would more closely correspond to educational and ecolinguistic categories. What we end up with is many overlapping sets. Viewed along the entire continuum of sociolinguistic rules we can probably come up with as many sets of grammars as there are native speakers of Arabic. Viewed along another dimension we have dialects which are grouped according to professions transcending national boundaries. This fits in very nicely with Cadora's theory of ecolinguistics. This model of "one person/one grammar" describes a very different and much more dynamic situation than what is described by the concept of diglossia.

This model of "inherent variability" is of course highly speculative and can only be proven after much research. However, it also fits in with what Versteegh is saying, because it models the mix of dialect and MSA (heavily favoring MSA) which is the final stage in his three phase process of pidginization, creolization and gradual decreolization. Since this process occurs over a large geographic area it is not surprising that there is a lot of regional variation.

According to Ferguson's diglossia theory diglossia will become unstable when literacy rates increase. Lacking precise statistics I will make the unsupported, but probably safe, claim that literacy rates have by and large greatly increased in Arab World since World War II. Thus, even according to Ferguson's original theory of diglossia, New Arabic's current situation can no longer be described as diglossia because literacy has become more widespread recently. Therefore we need to come up with a new model and some new terminology. From my perspective the model of "inherent variability" with a large set of accompanying socioloinguistic rules is as good a place to start as any.Return to Table of Contents

VI.2 Comparing Arabic with Latin and English

Languages other than Arabic have been subject to some of the same influences. For instance there is Latin. Latin, like Arabic, was the language of conquering armies and, like Arabic, the size of the conquered peoples' population was much larger than the size of the conquering armies. Eventually the conquered peoples became incorporated into the empire as citizens and adopted the language and customs of the conquerors. The relationship between the developing Romance languages and High Latin in the Middle Ages fits Ferguson's definition of diglossia almost perfectly, i.e. it possessed a large body of literature spanning many centuries combined with a very low literacy rate. There was a high variety of the language and there were low varieties which were used for most ordinary conversation. To my knowledge this is usually attributed to the fact that High Latin was frozen in time by the body of Latin literature at first and then by the Catholic Church. Because of this Vulgar Latin evolved out from under it, eventually becoming many separate geographically dispersed languages.

The New Spoken Arabic dialects have evolved unchecked while at the same time great care has been taken to keep Al-cArabiyya the same. This situation when viewed completely separately from any other consideration would seem very analogous to what happened to Latin as the Romance languages evolved. There is no arguing with the fact that left to their own devices languages do evolve. The Romance languages have evolved away from the highly synthetic Latin language. None of the Romance languages have preserved the case ending system, nor the mostly free word order that existed in Latin. To my knowledge folks don't argue about whether or not there was diglossia in Latin during the time of Caesar, or do they?

Another language which it would seem on the face of it to be undergoing some of the same influences as Arabic did is English. Today as I write this paper more than half of the people who use English in their day to day business are not native speakers. Admittedly most of these people have learned English in a tutored way. We don't have a situation where a large segment or even majority of the next generation is learning English from a mother who doesn't really know English very well. However, I am also running on the assumption that English is not changing very rapidly. English is part of the modern tradition which allows the written form of a given language to evolve more or less at the same rate as its spoken form. Now, I can still read English texts from 200 years ago, but if I go back farther than that it starts to sound odd to me. I can understand it, but I would never talk that way. I'm not sure I can trust my perception that English is not changing quickly.

I think it would be instructive to compare and contrast the spread of Latin and the development of the Romance languages with the spread and development of Arabic, for similarities and for differences. I also think it would be instructive to compare and contrast the development of post World War II Modern English with the spread of Arabic in the first 100 years of the Islamic conquests.Return to Table of Contents

VI.3 Putting Versteegh to the Test

Furthermore there is a lot of work to be done documenting the changes currently in progress with Arabic. Can we document the post World War II changes to Arabic? Are these changes uniform throughout the Arab World? It seems to me that by studying the current developments in Arabic we can test the soundness of the decreolization part of Versteegh's hypothesis. If his theory is true we would expect that the increased literacy rates in the Arab World are increasing the rate of decreolization in Arabic dialects. I would expect that according to Versteegh's theory, everything else being equal, the higher the literacy rate, the more pressure exists on that dialect in the direction of MSA. We should be able to study a dialect for the current direction of the changes taking place and determine if it is moving toward a local prestige dialect or toward MSA or toward both at once. If we suppose that Versteegh's theory is absolutely true then we would expect that the dialects are collapsing in the direction of MSA. If this is not the case we should then be able to either voice serious objections to this part of his theory or propose other influences which exert greater pressure on the dialects than MSA does. If we can observe the same trends in more than one dialect we should be able come to some definite conclusions.

On another front we can study the Arabic based creoles in the Sudan and perhaps in the horn of Africa to determine if they have developed according to Versteegh's model, although this seems less clear cut than testing the decreolization prediction. If we are lucky we might be able to observe first-hand the process of Arabic's continuing spread down the Nile in the Sudan, although this seems more like a long-term project spanning several generations.Return to Table of Contents

VI.4 Closing

In this paper I have briefly outlined some of the more prominent theories about the history and development of Arabic from the pre-Islamic period to the modern day. A lot of controversy surrounds this topic. The evidence is scanty and much of what is debated is speculative in nature. I have made a case that some of these theories are not mutually exclusive. In particular, we can explain a lot if we combine Versteegh's theory of pidginization, creolization and decreolization with Ferguson's KoinŽ II theory. I have outlined a few avenues for further research and have asked a few pointed questions.

I have outlined a proposal for a new way of looking at the relationship between the dialects and MSA which transcends the concept of diglossia. This idea is not originally mine, and I must give credit to Frederick Cadora who introduced me to it in the Fall of 1995 during the course of his "Phonology, Morphology and Syntax of Arabic" seminar at the University of Michigan. This model seems superior because it is a more holistic model and fits with the native speakers' perception that they only speak one language.

There is obviously a lot of work to be done in all of these varying areas of proposed research and investigation. This work should be done, and the debate should continue. Nevertheless, I will reiterate, in the final analysis we must admit that we may never know with certainty the truth about how Modern Arabic was formed from the earlier stages of the language.Return to Table of Contents

1. Blanc, H. 1964. Communal Dialects in Baghdad, pp. 160-171
2. Blau, J. 1965. The Emergence and Linguistic Background of Judeo-Arabic, pp. 1-50
3. Blau, J. 1966. A Grammar of Christian Arabic, pp. 19-58
4. Blau, J. 1977. "The Beginnings of the Arabic Diglossia: A Study of the Origins of Neo-Arabic," Afro-Asiatic Lingiustics 4: 175-202
5. Cadora, F. 1970. "Some Linguistic Concomitants of Contactual Factors of Urbanization," Anthropological Linguistics, 12: 10-19
6. Cadora, F. 1992. "Ecolinguistic Rules, Lexical Diffusion, and Historical Recapitulation," Bedouin, Rural and Urban Arabic: An Ecolinguistic Study, pp. 109-139
7. Cadora, F. 1992. "Linguistic Change and Ecolinguistics," Bedouin, Rural and Urban Arabic: An Ecolinguistic Study, pp. 1-35
8. Corriente, F. 1976. "From Old Arabic to Classical Arabic the Pre-Islamic KoinŽ: Some Notes on the Native Grammarians' Sources, Attitudes and Goals," Journal of Semitic Studies 21: 62-98
9. Diem, W. 1976 "Some Glimpses of the Rise and Early Development of Arabic Orthography," Orientalia 45: 251-261
10. DiPetro, R. and Selim, G. 1967. "The Language Situation in Arab Sicily," Linguistic Studies in Memory of R. S. Harrell, pp. 19-35
11. Ferguson, Ch. 1959. "Diglossia" Word 15: 325-337
12. Ferguson, Ch. 1959. "The Arabic KoinŽ," Language 35: 616-630
13. Ferguson, Ch. 1989. "Grammatical Agreement in Classical Arabic and the Modern Dialects: A Response to Pidginization Hypothesis," Al-Arabiyya, 22: 5-18
14. Hary, B. 1989. "Middle Arabic: Proposals for New Terminology," Al-Arabiyya 22:19-36
15. Kaye, A. 1970. "Modern Standard Arabic and the Colloquials," Lingua 24: 374-412
16. Moscati. S. 1964 "The Semitic Languges," An Introduction to Comparative Grammar of the Semitic Languages, pp. 3-21
17. Rabin, Ch. 1955. "The Beginnings of Classical Arabic," Islamica 4: 19-37
18. Versteegh, K. 1984. Pidginization and Creolization: The Case of Arabic
19. Ziadeh, F. 1986. "Prosody and the Initial Formation of Classical Arabic," JAOS 106.2: 333-338
20. Zwettler, M. 1978. "The Classical Arabiyya as the Language of an Oral Poetry," The Oral Tradition of Classical Arabic Poetry: Its Character Implications, pp. 97-172
Return to Diglossia Page
Return Home الرجوع إلى الصفحة الرئيسية