Arabia before the first letter descended
Before a single word of the Quran was written down in this history, the pen already existed. Before there was a Messenger ﷺ to dictate, and before there was a scribe to record, the tool with which recording would be done already existed. That sounds self-evident, but it is the starting point of a battle that would rage for a century and a half in the studies of Europe, and it is the reason that this entire timeline does not begin at a cave in the year 610, but on a rock in the year 512.
It is a tempting thought, and it has been cherished for centuries, that the Arabia of before Islam was a land without letters. A desert of sound and memory, where the poet recited and the caravan remembered, but where no one wrote. In that image the Arabs were a people of ears and tongues, not of eyes and hands. They preserved their history in verses that passed from mouth to mouth, their genealogies in the heads of the old men, their honour in the sharpness of a mocking poem. Paper and parchment, so it was supposed, were the instruments of the civilized neighbouring peoples: the Byzantines with their Greek, the churches of Syria with their Aramaic, the rabbis with their Hebrew. The Arab himself would only have acquired a script when he needed it, and that moment would coincide, or almost coincide, with the coming of revelation.
If that were true, it would have consequences that reach far beyond a footnote about palaeography. For if the Arabs at the beginning of the seventh century could barely write, then the Quran could not possibly have been recorded as the tradition maintains: immediately, carefully, by dozens of trained scribes, on parchment and leather and shoulder blades. Then the written Quran would have to be a later construction, raised up in a time when the script was indeed mature, generations after the Prophet ﷺ. Thus, from the absence of an alphabet, an argument was forged against the reliability of the entire Book.
This article opens the timeline of the compilation of the Quran, and it opens it deliberately not at the first revelation but before it. To understand how the Quran was written down, one must first know that the script in which it was written down already existed. Not as a raw, newborn sign-language, but as a mature, dated, recognizable Arabic script, which a century and more before the Prophet ﷺ was carved into stone and which stands so close to the script of the first generation of Muslims that the trained eye can scarcely draw a boundary between them. The Quran was not written in a script that was invented for it. It was written in a script that it inherited.
There is something almost inverted in this order. We are accustomed to letting the history of Islam begin with revelation, and rightly so, for that is where the message begins. But the question of the script is a question that lies before the message, just as the question of the parchment lies before the question of the ink. A message can only be written down once there is a script to write it in, and a script is not born in a single night. It ripens over centuries, from hand to hand, from stone to stone, from generation to generation. Whoever wishes to know whether the Quran could have been recorded as the tradition describes must therefore first look at the centuries that preceded the revelation, and at what the hands of the forefathers had already left behind there.
That is the subject of this first part. We descend into the sixth century, into the time before the birth of the Prophet ﷺ, and we look at three stones. They are not spectacular stones. There is no gold upon them, no name of a king whom history has remembered, no battle that changed the world. There are only letters upon them, Arabic letters, with a date beside them. But it is precisely those sober, dated letters that carry a weight that few golden monuments can equal. They prove that the script existed before the revelation came, and with that they remove the first block from a wall that for a century and a half was raised against the reliability of the Quran.
The accusation of Mingana
Every refutation needs an accusation, and in this case the accusation has a name. Alphonse Mingana, the Chaldean priest whose manuscript collection would later end up in Birmingham and whose folios, ironically enough, would one day prove to preserve one of the oldest material Qurans in the world, formulated at the beginning of the twentieth century a thesis that would set the tone for an entire school of critical Quran scholarship. His premise was sharp and simple: the early Arabs did not possess their own developed alphabet capable of recording a text like the Quran. What they had in the way of script was, in his eyes, a borrowed and defective instrument, derived from the Syriac estrangelo or even from Hebrew, and too undeveloped to bear what the tradition made it bear.
It was not a passing remark. It was a foundation. Whoever assumes that there was no mature Arabic alphabet in the time of the Prophet ﷺ has opened the door to the conclusion that the written Quran must be a product of later times. The assumption about the script thus became the first stone of a building that, floor after floor, would contest the reliability of the Quranic text. Later scholars would build upon it with theories of a Quran that was only compiled in the second or third century after the hijra, but they all built upon the same ground floor: the Arabia of the Prophet ﷺ was supposed to have been a land without mature script.
One must understand why this type of argument was so attractive to whoever wished to overturn the traditional history. The tradition of the Muslims is rich, detailed, and internally coherent; it names the names of scribes, the materials on which writing was done, the occasions at which dictation took place. To contest that tradition with counter-traditions is difficult, for the sources stand on the side of tradition. But whoever attacks the premise about the script need not refute the tradition piece by piece; he need only claim that it is impossible. If there was no alphabet, then all the accounts of scribes and dictations cannot be true, however detailed they may be. It is an attack not on the branches but on the root. One premise, if true, would let an entire forest of traditions topple at once.
That is precisely why the refutation of it is so decisive. An attack on the root is turned back at the root. When one demonstrates that the mature Arabic alphabet did indeed exist, and amply before the Prophet ﷺ, then not one tradition falls but the entire attack falls, and the traditions that it wished to overturn remain standing where they stood. The stones of this article are therefore not merely a nice palaeographic curiosity. They are the defence of the root of the entire written Quranic tradition.
Against that premise Mohammad Mustafa Azami ﵀ directs a chapter of his standard work on the history of the Quranic text. His approach is not rhetorical but material. He does not advance an argument of which the opponent can say that it is a pious belief; he advances stones. Dated stones, with Arabic letters upon them, from a time that makes the accusation impossible. Azami states his intention explicitly: he will present a number of dated and highly developed inscriptions that clearly demonstrate the contrary, inscriptions from the sixth century after Christ that very closely approach the Arabic palaeography that was used in the first century after the hijra, the seventh century after Christ. The formulation is precise, and the precision matters: he does not say that the sixth-century inscriptions are identical to the later script, but that they approach it. That distinction we shall hold onto, for the strength of the argument lies precisely in its restraint.
An old language with an old alphabet
To understand why Azami is so certain, one must take a step back into the history of the Arabic script itself. The script that the Muslims used in the seventh century did not come out of nothing. It had ancestors, and those ancestors run back to a people whom most people do not connect with Arabic: the Nabataeans.
The Nabataeans were the builders of Petra, the caravan traders of the north of the Arabian Peninsula, whose kingdom extended from before our era until the year 105 after Christ, when Rome annexed it. For a long time it was assumed that their language and their script were Aramaic. Azami argues, following historians who have reexamined the question, that this is a misunderstanding that rests on a mistaken name. The Nabataeans spoke Arabic, and their script was an early form of the Arabic script. That one called it “Nabataean,” Azami says, is no different than when one speaks of “Egyptian Arabic” or “American English”: a regional name for something that is essentially Arabic. The Arabic language and the Arabic script, in their early forms, gave rise to the Nabataean, and in their origin probably preceded the Syriac.
That is an inversion of the image that Mingana supposed. Where the accusation portrayed Arabic as a late borrower of the Syriac script, Azami places Arabic precisely before Syriac in the development. The letters that the Muslims would use in Medina had a lineage that reached back centuries, straight through the Nabataean, into a time when Arabic was already a written language.
Azami reinforces this with an argument that he draws from an opponent. Gruendler, who studied the Arabic script, observed that certain sounds of Arabic are not represented in Aramaic, and concluded that the writers of the Nabataean texts spoke Arabic, because otherwise they could not have rendered those unusual sounds in their spelling. Put differently: the Nabataean language and the Nabataean script were in fact a form of Arabic. It is a concession from an unsuspected quarter, and Azami knows how to appreciate it. For if even the scholar who treated the Arabic script as a late and derivative phenomenon must concede that the Nabataeans spoke Arabic and that their script carried those Arabic sounds, then the notion that Arabic would only have acquired a script around the time of the Prophet ﷺ is untenable from within.
There runs a still older line through this history, which Azami likewise touches upon and which anchors the roots of the Arabic script in the patriarchs. Ismail ﵇, the eldest son of Ibrahim ﵇, grew up at Mecca among the tribe of Jurhum and married within that tribe. The Jurhum spoke Arabic, and so, as Azami reasons, Ismail ﵇ spoke it too. His descendants, who would eventually found the Nabataean kingdom, took the Arabic language and the Arabic alphabet with them when they moved north. In this way the lineage of the script is traced back, not to a late borrowing from the neighbours, but to an inheritance that travelled along with the descendants of Ibrahim ﵇ and Ismail ﵇. The Arabic script in this view is no borrowed garment but a family possession, handed down from generation to generation, long before it would receive its highest commission.
And there is an older witness still. The inscription of Imru’ al-Kais at Namarah, about a hundred kilometres to the southwest of Bosra, is dated to approximately 328 after Christ. There has been learned dispute over whether it should be called Nabataean or Arabic, whereby a scholar such as Gruendler regards it as Nabataean and others, among them Cantineau and Abbott, treat it as Arabic. But Azami points to something that rises above that quarrel: the dividing line between what one calls Arabic and what one calls Nabataean is extraordinarily thin, and the Namarah inscription now counts as the oldest known dated Arabic inscription. More than two hundred years before the Prophet ﷺ, then, a dated Arabic text already stood carved in stone. The question is no longer whether the Arabs before Islam could write, but how mature that script already was by the time the revelation came.
To that question Azami gives an answer in three stones. They form the heart of this article, and together they form the refutation.
Zabad, approximately 512 after Christ
The first stone lies in the north of Syria, at a place called Zabad. It is a pre-Islamic inscription, and it is set not in one language but in three: Arabic, Greek, and Syriac, side by side on the same surface. Azami dates it to approximately 512 after Christ, and he draws it from the work of the palaeographer al-Munajjid.
Let the image sink in for a moment. Approximately a century before the first revelation descended, someone in the north of Syria hewed into stone a text in which the Arabic script stood shoulder to shoulder with the Greek of the Byzantines and the Syriac of the churches. The Arabic was no second-rate guest there, no scribble in the margin of richer languages. It stood as an equal between the two script cultures that dominated the surroundings. A script that one dares to place beside the Greek and the Syriac on a single memorial stone is not a script that one is ashamed to show. It is a script that one trusts.
The dating of approximately 512 is here no ornamental detail but the core of the argument. It lies almost a full century before the year 610, in which the Prophet ﷺ received his first words in the cave of Hira. Whoever claims that the Arabs in 610 did not yet possess a proper alphabet must explain how in 512 there could already stand a trilingual inscription in which that alphabet was carved with a steady hand beside the Greek and the Syriac. The stone of Zabad is older than the accusation that tries to deny it.
The word “approximately” earns its place here. Azami does not write that the inscription was made in exactly 512, but that it is dated to approximately that year. Such datings of early inscriptions always carry a margin; a temple calendar, a regnal year, an indiction, all must be converted and weighed. But even with a generous margin the stone of Zabad remains standing where it stands: in the sixth century, well before Islam, with a full-grown Arabic script upon its surface.
Jabal Usays, approximately 528 after Christ
The second stone lies a hundred and five kilometres to the southeast of Damascus, at a place that appears in the sources under more than one name. Azami spells it as Jabal Asis. In other works one finds it back as Jabal Usays or as Jabal Says. It is important to set this straight at once, for the different spellings can give the impression that different sites are meant, whereas it concerns one and the same place. Jabal Usays, Jabal Asis, and Jabal Says are three spellings of the same name, of the same mountain, of the same inscription. Where this article cites the site under the name Jabal Usays, and where it retains Azami’s own spelling Jabal Asis in the direct reference, it always concerns the same point on the map.
The inscription is again pre-Islamic and again Arabic, and it is dated to approximately 528 after Christ. Azami draws the dating from Hamidullah, in his collection of the earliest diplomatic documents. We are now sixteen years further than Zabad, and still more than eighty years before the first revelation. The stone lies not far from Damascus, in the Syrian desert that the Arab tribes crossed on their trade routes between the peninsula and the north. This is no script that hides itself in a remote corner; it is a script that is used along the roads of trade, in the region where Arabs, Byzantines, and Syrian Christians met one another daily.
Two stones, two datings, two sites, and each time the same Arabic script, each time well before the Prophet ﷺ. The pattern begins to take shape. The claim that the Arabs did not possess a mature alphabet is refuted not with one chance find, but with a series. And the series is not yet finished.
Harran, approximately 568 after Christ
The third stone bears the name Harran, a place that the reader of the older history knows as an ancient centre in the north of Mesopotamia and Syria. Here too it concerns a pre-Islamic Arabic inscription, and Azami dates it to approximately 568 after Christ, again on the authority of al-Munajjid.
The year 568 lies only a few years removed from the birth of the Prophet ﷺ, which the tradition places near the year 570, in the Year of the Elephant. That is a striking coincidence. In the decade in which the man would be born to whom the Quran would be revealed, the Arabic script already stood so firm that one erected dated monuments with it. When the Prophet ﷺ walked as a child through the streets of Mecca, the script in which his revelation would one day be recorded was already a generation old in stone.
With this third inscription closes the series that Azami builds up from the sixth century toward the Islamic period. Zabad around 512, Jabal Usays around 528, Harran around 568: three dated points, spread over more than half a century, all three in the north of the Arabic world, all three in an Arabic script that is not raw or hesitant but developed. They are not the scratches of someone who is just learning to write. They are inscriptions of a script culture that masters its tools.
Very close, not identical
Here the careful reader must be on guard, for it is precisely on this point that it is easy to exaggerate, and the exaggeration would weaken the argument instead of strengthening it. Azami does not claim that the script of Zabad, Jabal Usays, and Harran is identical to the script of the first generation of Muslims. He claims that it very closely approaches that script.
The distinction is fine but essential. A script develops, always and everywhere. The letters change in form, the connections become more fluid, the hand gains in ease. Between the sixth-century inscriptions and the masahif of the first century after the hijra lies a stretch of that development. Whoever would claim that the stone of Zabad looks exactly like a leaf from an early Quran would do violence to the historical reality and would expose himself to the justified criticism of any palaeographer. Azami does not do that. His claim is more modest and thereby stronger: the sixth-century Arabic palaeography already finds itself so close to that of the first century after the hijra that no room remains for the notion that the mature Arabic alphabet would only have arisen after the Prophet ﷺ.
That is precisely what the accusation needed in order to stay upright, and that is precisely what the stones deny. Mingana needed a leap: an Arabia without mature script in the seventh century, in which the written Quran was impossible and therefore must have been constructed later. The inscriptions remove that leap. The script that characterizes the first century after the hijra lies in the continuation of what already stood in stone in the sixth century; it is the same tradition, a small stretch further in its growth. There is no chasm in which a later invention could be lodged.
And then comes the conclusion toward which the chapter works. These inscriptions, writes Azami, sufficiently refute the premise of the reverend Mingana regarding the early Arabic alphabet. It is a sober sentence, without triumph, and precisely for that reason convincing. The evidence has been laid down, the accusation has been weighed, and the accusation falls. The Arabs of before Islam possessed their own, recognizable, dated Arabic alphabet. Whoever claims otherwise must explain the stones away, and the stones do not let themselves be explained away.
The script keeps growing
That the Arabic alphabet was already mature before Islam does not mean that it stood still. A living script is like a living language: it keeps developing as long as people use it. And when the young Islamic state extended from Azerbaijan and Armenia in the north to Yemen in the south, from Libya and Egypt in the west to Iran in the east, and all those far corners received their instructions from Medina in Arabic, that development accelerated sharply. There arose a rich layering of script styles, whereby an angular or non-fluid variant developed at a very early stage alongside the Hijazi. This angular style has received in modern scholarship a name that has itself become a source of confusion: Kufic. Azami expressly enters an objection to lumping all angular script styles together under that one word, for the earliest scholars distinguished a dozen or more styles, of which the actual Kufic was only one.
Around precisely that angular script a second accusation has been built, related to that of Mingana but placed one layer later. Where Mingana wished to explain away the script before Islam, others wished to date the angular script late, after Islam: every Quran in Kufic script, so they claimed, could only stem from the second or third century after the hijra, and thus be suspiciously late. It is the same fallacy, one floor higher: an assumption about the script deployed to date the text late, and again to be refuted with dated stones from the first century itself.
That refutation, however, deserves its own treatment, and it receives it further on in this series. When we later dwell upon a Kufic Quran fragment that natural science itself, outside every Muslim hand, has dated to the seventh century, the dated first-century examples on which Azami relies come extensively to the fore. For this article, which is about the time before Islam, the observation suffices that the script kept developing and branching after the revelation, without ever showing the chasm in which a later invention could be hidden. The three stones of the sixth century close off the root; the first-century stones, when we reach them, close off the branch.
Where the stones lie
A word about the place that these three sites occupy on the map, for the reader who follows this timeline on the map deserves to know what he sees and what he does not see.
Zabad lies in the north of Syria, to the southeast of present-day Aleppo. Jabal Usays lies a hundred and five kilometres to the southeast of Damascus, in the basalt desert that stretches toward the interior. Harran lies in the border region of the north of Syria and Mesopotamia, on the old route between the rivers. The coordinates with which these points are placed on the map are modern estimates; Azami himself gives no coordinates, and the exact location of an inscription in stone always leaves some latitude. The reader should therefore read the pins on the map as approximations of the region, not as geodetically fixed points. What stands firm is not the precise latitude, but the region and the dating: the north of the Arabic world, the sixth century, a mature Arabic script.
It is no coincidence that all three stones lie in the north. There, along the trade routes and at the edges of the Byzantine and Syrian spheres of influence, the Arab tribes met the script cultures of their neighbours, and there the most was preserved in stone. The peninsula itself, with its dry air and its meagre stone, has left us fewer of these kinds of dated monuments, but that does not mean that the script was unknown there. Trade brought it along, the caravan carried it, the poet and the merchant knew it. Mecca, the trading city par excellence, was no island of illiteracy in a sea of literacy; it lay in the middle of the stream.
What the refutation carries and what it does not carry
As with every strong argument, here too one must distinguish sharply what it proves and what it does not prove, for the strength of a refutation lies in its honesty about its own limits.
What the inscriptions prove is this: a mature, dated Arabic alphabet existed well before the Prophet ﷺ. The premise that the early Arabs did not possess their own developed script is thereby refuted. The foundation beneath the later theory, that the written Quran would be a construction from after the time of the Prophet ﷺ because the script was then not yet mature, is knocked away. If the script was already mature in the sixth century, then the written recording of the revelation in the seventh century cannot be rejected on the ground of the impossibility of the script. The way is clear for precisely that which the tradition describes: scribes who, at the direction of the Prophet ﷺ, record what is revealed.
What the inscriptions do not prove is equally important to say. They do not in themselves prove that the Quran was written down; that is a separate question, with its own evidence, which comes to the fore in the following articles. They do not prove that every Arab could read and write; literacy was in the sixth and seventh centuries everywhere in the world the possession of a minority, and Arabia formed no exception to that. They also do not prove that the script of 512 looked the same as that of 650; we have seen that Azami emphasizes precisely the contrary, that the script developed and that the sixth-century form only approaches the later.
What the inscriptions do do is pull the floor out from under a certain way of reasoning. They make it impossible to say that the Arabs could not write down the revelation for lack of a script. The script was there. It lay on the rocks of Zabad, of Jabal Usays, of Harran, dated and recognizable, well before the first revelation descended. The question that remains is not whether one could write down the Quran, but how carefully one did it. And to that question the rest of this history gives an answer.
There is one more detail that sharpens the image of a literate environment around the Prophet ﷺ, and that returns extensively later in this history. When an orientalist cast doubt on the authenticity of the preserved letters of the Prophet ﷺ with the argument that the handwritings differ among themselves and thus could not come from the same state, Azami dismissed that as nonsense, with a reason that says everything about the literacy of the early umma: the Prophet ﷺ had more than sixty scribes, and to expect that their handwritings would resemble one another is absurd. That number is no side issue. A community that could deploy sixty scribes for the correspondence and the revelation of its leader is no community that struggled with a newborn alphabet. It possessed a corps of scribes, a practised craft, hands that were familiar with the pen. The script that stood on the rocks of the sixth century had grown, by the seventh century, into a daily tool in the hands of dozens.
The inscriptions and the scribes together point in the same direction. From the dated stone at Zabad to the sixty scribes in Medina runs one unbroken line: a script culture that was mature before the revelation came and that stood ready when the revelation came. The accusation of a scriptless Arabia thereby hangs in the air, without ground beneath its feet. Where it supposed an emptiness, there stands a tradition.
The script that waited
It is useful, at the end of this first episode, to consider what all of this means for the way in which the Quran took its place in history.
The Quran was not sent down into a cultural vacuum. It was sent down among a people that possessed a language of unequalled richness, whose poetry touched the most sensitive strings of human eloquence, and that, so it appears from the stones, also possessed a script to record that language. The challenge that the Quran posed to the Arabs was therefore no challenge to a primitive people without the means to answer it. It was a challenge to the masters of the language, in their own language, in their own script. That only sharpens the nature of the Quran. The Book challenged not the ignorant but the knowledgeable, not the speechless but the eloquent.
And the script that the Muslims used to record the revelation was not something that they had to invent hastily while the verses descended. It was an instrument that lay ready, inherited from generations who had hewn into stone before them. When the Prophet ﷺ called his scribes to record what was revealed, they reached for a tool that had already lain for centuries in the hands of their forefathers. The Arabic alphabet stood, as a practised tool stands in the workshop of a craftsman, ready for the work that it had to do. The revelation did not need to create the script; it needed only to use it.
This is why the timeline of the compilation begins on a rock and not in a cave. Before anything was revealed, it had to stand firm that there was something to write it in. The stones of Zabad, Jabal Usays, and Harran fix that certainty in the hardest material that history knows. They say, in a language that needs no interpreter: the script already existed.
And so the history can now proceed to the moment at which that script would find its highest destination. Approximately a century after the stone of Zabad, in a cave above Mecca, on a night in Ramadan, an angel would say to an unlettered man: read. And that same man, who himself could neither read nor write, would gather scribes around him who indeed could, and they would reach for the pen that their forefathers had left them. The script that had lain waiting on the rocks of Syria would become the script in which the final revelation to mankind was recorded, letter after letter, verse after verse, surah after surah. In the next part we descend that century, to the cave of Hira, to the first five verses, and to the promise enclosed within them: that the Lord teaches man with the pen.
والله أعلم