The Paper Before the Book

history of paper in the Islamic world: typographic cover reading "A civilisation written on recovered cloth"

How a material changed the scale of knowledge across the Islamic world

The history of paper in the Islamic world does not begin with a battle. Paper reached these lands slowly, across Central Asia, over the course of the eighth century. It was not carried there by Chinese papermakers captured at the Battle of Talas in 751: of the letters recovered from the fortress of Mount Mugh, written before August 722, thirteen are already on paper. What followed Talas was not a transmission. It was an adaptation, and it took three hundred years.

Before there was a book, there was a surface.

Before libraries, copyists, book markets and monumental Qur’ans, there had to be something on which words could remain.

For centuries, that surface had been many things: parchment, papyrus, wood, leather.

Then paper entered the story.

Its history in the Islamic world is usually told through a single dramatic scene. In 751, Muslim forces defeated a Chinese army at Talas, in Central Asia. Chinese papermakers were supposedly captured. They were taken to Samarkand. They revealed the secret.

It is a memorable story.

It is also far too simple.

Jonathan M. Bloom, whose Paper Before Print remains the major study of paper in the Islamic world, is widely reported to have judged the Talas anecdote unlikely to be factual. His argument, as it is summarised, is not that the battle did not happen — it did — but that Chinese and Central Asian papermaking relied on different materials: bast fibres on one side, rags on the other. A captured craftsman does not explain a different craft.

And the evidence from the ground is older than the battle.

Thirteen letters, written before the battle

In the early 1930s, in a ruined fortress on Mount Mugh in what is now Tajikistan, an archive was found by accident. It holds seventy-eight texts, mostly Sogdian, with one Arabic document, one Old Turkic and a few Chinese fragments. Among them are thirty-five letters.

Thirteen of those letters are written on paper. Thirteen are on parchment. Eight are on wooden sticks.

The dated documents run from before 708 to the summer of 722. They belong to Dēwāshtīch, called the last king of Sogdiana, who fled up the Zeravshan after Panjikent was sacked, was captured at Mount Mugh in August 722, and was executed.

Twenty-nine years before Talas, in a mountain fortress in Central Asia, a ruler and his administration were already writing on paper — alongside parchment, alongside wood, as one material among others.

That is the whole correction, and it is a count rather than an opinion.

Paper was already there. What happened next was not a single moment of transmission. It was a process of adoption, adaptation and scale.

Before Talas

The temptation to tell history through beginnings is understandable.

We like firsts.

The first paper mill. The first book. The first city to adopt a technique. The first person to introduce it.

But materials rarely obey such clean timelines.

Paper had originated in China centuries earlier, then travelled gradually across regions connected by trade, migration and political change. Central Asia was one of those regions, and Samarkand, placed on major commercial routes, became one of the most important places in this story.

But it was not a waiting room between China and the Islamic world.

Local traditions evolved. Different fibres were used. Different techniques appeared. The material itself changed.

The shift most often singled out is the extensive use of textile fibres — especially old rags — in place of the specialised plant sources of earlier Chinese practice. I have not been able to trace that account to a technical study, and it is given here as it is generally told; but if it holds, it changes the economics of the material entirely.

Worn cloth, textile waste and old ropes could become raw material.

Something already used could become something on which new knowledge could be written.

That transformation is easy to overlook. It is also one of the reasons paper could become powerful.

The history of paper is not only the history of an invention. It is the history of a material made reproducible.

How do you make a sheet of paper?

It begins with destruction.

Old textiles are collected. They are cleaned. Broken down. Beaten until their fibres separate.

The fibres are suspended in water, creating a pulp.

A mould is lowered into that mixture. When it is lifted, a thin layer of fibres remains across its surface. Water drains away. The fibres settle together.

A sheet begins to exist.

The process is repetitive, physical and precise.

One sheet. Then another. Then another.

Accounts of the early Islamic tradition describe different moulds over time, including floating moulds and later two-part moulds that made it easier to lift a newly formed wet sheet and continue production. The same accounts describe what followed: a raw sheet was not yet a writing surface. It could be sized with starch drawn from wheat or rice, then scraped, polished and burnished, with glass or agate used to smooth it. These descriptions are widely repeated; I have not verified them against a published technical study, and they are reported here as such.

A sheet of paper was therefore not simply produced.

It was finished. Prepared for ink. Prepared for the hand. Prepared for the possibility that something written on it might need to survive for centuries.

The size of a sheet

There is something unexpectedly human about the dimensions of early paper.

A figure commonly given holds that sheets rarely exceeded roughly 45 by 65 centimetres. I could not trace it to a published measurement, so take it as an approximation that circulates rather than a number that has been established.

The explanation offered for it, however, is worth keeping.

A single papermaker had to handle the mould. The possible size of the page was therefore limited by the body of the person making it. By reach. By balance. By the width of two arms holding a frame over water.

Long before page formats were standardised by machinery, the sheet carried the measure of the human body.

There were exceptions. Larger sheets existed, and as papermaking developed, particularly in Iran, larger formats would help make monumental Qur’ans and large illustrated manuscripts possible.

But the principle remains beautiful in its simplicity. The page began as something a person had to be able to hold.

From Samarkand to Baghdad

By the late eighth century, paper had become increasingly visible in the Abbasid world, and Baghdad became one of the great centres of its use.

Tradition associates the expansion of papermaking there with the reign of Hārūn al-Rashīd and with the vizier al-Faḍl ibn Yaḥyā. It is an attribution repeated for centuries rather than a documented fact, and it is worth holding at that distance — especially in an article whose first argument is that a single founding scene rarely explains a material.

What is less disputed is the direction of travel. By around the beginning of the ninth century, paper was established enough to be used in administration.

That point matters.

Governments do not adopt a writing material because it is interesting. They adopt it when it becomes practical. Available. Reliable. Useful at scale.

Paper entered the machinery of administration. Then the machinery of scholarship. Then everyday commercial life.

A Greek Christian manuscript is reported to have been copied on Arabic paper around the beginning of the ninth century — a detail I have not been able to source precisely, but one worth naming, because it says the material was already moving between languages and religious communities.

The oldest dated book, and the one that may displace it

The oldest dated book in Arabic script on paper that has been published and studied is a copy of the Gharīb al-ḥadīth of Abū ʿUbayd al-Qāsim b. Sallām, held at Leiden University Library as Legatum Warnerianum Or. 298. Its colophon gives Dhū al-Qaʿda 252, which sources convert as late 866 or late 867. It bears no indication of where it was copied.

It is generally called the second oldest, not the first.

A manuscript dated 848 was found by accident in the regional library of Alexandria. It awaits publication.

So the honest sentence is not that we know which book came first. It is that the oldest one we can read is in Leiden, that an older one exists in Alexandria, and that until it is published the question stays open.

Cheap enough to wrap goods

Perhaps the most revealing moment in the history of paper does not come from a royal library. It comes from a market.

In the eleventh century, the Persian traveller Nāṣir-i Khusraw is reported to have described merchants in Cairo wrapping goods in paper. The detail is often quoted; I was not able to verify the passage in a reliable edition of his Safarnāma, and it is repeated here as a report rather than a citation.

If it is right, it is extraordinary precisely because it is so ordinary.

Paper had become cheap enough to use as packaging.

Think about what that means. A material that once represented specialised knowledge and long-distance technical transmission had become common enough to protect merchandise in a bazaar.

This is what scale looks like. Not a proclamation. Not a grand invention story. A shopkeeper wrapping an object.

When a material becomes mundane, it has usually already changed the world.

Kairouan, Fès, al-Andalus

From the eastern Islamic world, papermaking spread and adapted to local conditions. Accounts place production at Kairouan by the tenth century, name Egypt as an important producer, describe Fès as a significant centre, and give Xàtiva in al-Andalus a reputation for its paper. By the eleventh century, production in the western Islamic world is described as firmly established, and from there the material entered European networks. These are the dates as they are generally given; I have not traced them individually.

The story is sometimes compressed into another neat formula:

China invented paper. Muslims transmitted it. Europe perfected it.

This is no better than the Talas legend. It reduces centuries of adaptation to a relay race.

The Islamic world did not merely carry paper westward. Papermakers altered raw materials. They developed manufacturing practices. They expanded markets. They integrated paper into administration, scholarship, religion, commerce and artistic production.

Transmission happened. But transformation happened too. And transformation is the more interesting story.

When the Qur’an moved onto paper

For early Qur’anic manuscripts, parchment had been a dominant material. Paper changed what was possible. It was less costly. It could be produced in larger quantities. It encouraged new relationships between script, format and page.

This did not mean parchment disappeared overnight. Materials overlap. Practices coexist. Prestige survives technical change.

But over time, paper became increasingly central to the manuscript cultures of Islamic societies. And the consequences went far beyond Qur’anic copying.

Legal documents. Letters. Scientific works and the instruments they described. Poetry. Philosophy. Medical texts. Accounts. Administrative records. Architectural drawings. Pattern books.

The page became a place where knowledge could be copied, revised, carried and accumulated.

Knowledge no longer depended only on memory, oral transmission or expensive writing surfaces.

It could multiply.

Not infinitely. Not yet. That would require print.

But enough to change scale.

Before the wall, there was a sheet

There is another consequence of paper that is easy to miss. Architecture and ornament depended on it too.

Geometric patterns could be drawn. Proportions tested. Designs transmitted. Motifs preserved and adapted.

A pattern carved into stone or laid into tile may look permanent once it reaches a building. But before the wall, there may have been a page. Before the dome, a drawing. Before the ornament, a line on paper.

The material became a space for thinking before building.

The manuscript and the monument were not entirely separate worlds. A sheet could stand between them.

The paper seller becomes something more

As paper became more available, new professions grew around it. Among them was the warrāq.

The word covers a stationer, a copyist, a bookseller, someone working inside the wider ecosystem of manuscripts. A warrāq might copy texts. Sell paper. Commission manuscripts. Correct them. Bind them. Circulate them. He might also sell the instruments that wrote on them.

His shop could become more than a shop. It could become a place where texts were found, copied, discussed and transmitted.

One warrāq, and the measure of what paper had done

The most famous of them carried the trade in his name.

Abū’l-Faraj Muḥammad b. Abī Yaʿqūb Isḥāq al-Warrāq al-Nadīm was born in Baghdad around 320 of the Hijra, 932 of the common era. He earned his living copying and selling books, as his father had done before him.

In 377 of the Hijra, 987 of the common era, he began the Fihrist: a catalogue of the books, lecture notebooks and papers available in Arabic in his time. Ten discourses, from revealed scripture to grammar, history, poetry, theology, jurisprudence, philosophy, medicine, legends and alchemy. He left blank spaces for later additions.

Here is the measure the story needs.

Within roughly two centuries of paper becoming ordinary in the offices of Baghdad, a paper seller of that city could sit down and attempt to list everything written in his language.

Not an estimate. Not a contested library figure. A book, dated, that still exists.

Because once paper becomes abundant enough, the question changes. It is no longer how we make the sheet. It becomes: who writes on it? Who copies the text? Who pays for it? Who sells it? Who reads it? Who preserves it? Who carries it from one city to another?

The material creates the possibility. People create the culture.

Paper did not create knowledge

It would be easy to overstate the argument.

Paper did not create Islamic scholarship. It did not create the Qur’an. It did not create translation, mathematics, medicine, philosophy or architecture. Muslim societies were intellectually rich before paper became widespread.

A civilisation is never produced by one material.

But materials change what a civilisation can do at scale.

Paper lowered the cost of writing. It expanded the possible volume of copying. It supported administration. It widened the circulation of texts. It gave scholars, scribes, merchants and institutions a surface that could be produced more readily than parchment. It made books easier to multiply long before printing made them easier to reproduce.

That distinction matters.

Paper did not invent knowledge. It changed its reach.

A civilisation written on recovered cloth

There is something quietly striking about the material itself.

Old cloth. Worn fibres. Used ropes. Things that had already served one purpose.

Broken down. Beaten into pulp. Lifted from water. Pressed. Dried. Polished. Then written upon.

A discarded textile could become a legal document. A scientific diagram. A page of poetry. A merchant’s account. A Qur’anic manuscript. A sheet in a scholar’s hand.

The history of paper in the Islamic world is therefore not a story of transmission from East to West. It is the story of what happens when a civilisation takes a material and makes it useful to its own intellectual, administrative, artistic and spiritual life.

Paper travelled. But it did not remain unchanged. Nor did the societies that adopted it.

And once enough sheets existed, another world could emerge around them. The copyist. The bookseller. The binder. The manuscript market. The library. The warrāq.

Because before the history of printing could begin, there first had to be a world in which books were already everywhere.

Questions readers ask

What does the history of paper in the Islamic world actually show?

That a material arrived gradually and was then remade. Paper was in use in Central Asia before 722. Over the following centuries it was altered — different fibres, different moulds, different finishing — until it was cheap enough to wrap vegetables and abundant enough that a bookseller could try to catalogue everything written in Arabic. The story is not transmission. It is adaptation.

Did Chinese prisoners captured at Talas bring papermaking to the Muslim world?

The battle happened, in 751. The explanation does not hold. Paper was already being used in Central Asia decades earlier: thirteen of the thirty-five letters in the Mount Mugh archive, which ends in August 722, are written on paper. Chinese and Central Asian papermaking also relied on different raw materials, which makes a direct transfer of technique unlikely.

When does paper first appear in Central Asia?

Before 708 at the latest, on the evidence of the dated documents from Mount Mugh, which run from before 708 to the summer of 722. The archive was found by accident in the early 1930s in a ruined fortress in what is now Tajikistan.

What was Islamic paper made of?

Textile fibre, as it is generally described: old rags, worn cloth, used ropes, beaten into pulp and lifted from water on a mould. That is the difference most often drawn with earlier Chinese practice, which relied on bast fibres from specific plants.

What is the oldest dated Arabic book written on paper?

The oldest one that has been published is a copy of the Gharīb al-ḥadīth of Abū ʿUbayd al-Qāsim b. Sallām, at Leiden University Library, Or. 298, dated Dhū al-Qaʿda 252. It is usually called the second oldest: a manuscript dated 848, found in the regional library of Alexandria, is older but has not yet been published.

What is a warrāq?

A paper seller, and more than that: a copyist, stationer, bookseller and sometimes publisher. Ibn al-Nadīm, who compiled the Fihrist in 987, carried the word in his name and earned his living that way, as his father had.

Did paper cause the flowering of Islamic scholarship?

No. Muslim societies were intellectually rich before paper spread, and no civilisation is produced by a material. What paper changed was scale: the cost of writing, the volume of copying, the reach of a text. It made books easier to multiply long before printing made them easier to reproduce.

Honest reservations

This article separates what is documented from what is reported. The following points are reported, and are not established here.

  • The phrase attributed to Jonathan M. Bloom on the Talas story circulates as unlikely to be factual. I found it only in a secondary summary, not in Paper Before Print itself. It should be checked against the book, with a page.
  • The colophon of Leiden Or. 298 gives Dhū al-Qaʿda 252. Sources convert this as 866 or as 867. The conversion should be taken from the library’s own catalogue.
  • The manuscript dated 848 in Alexandria is unpublished. The ranking of Or. 298 as second oldest rests on it.
  • The figure of roughly 45 by 65 centimetres for a sheet is widely repeated. I could not trace it to a published measurement.
  • The descriptions of floating and two-part moulds, of sizing with wheat or rice starch, and of polishing with glass or agate, are taken from general accounts and not from a technical study.
  • The association of papermaking in Baghdad with Hārūn al-Rashīd and al-Faḍl ibn Yaḥyā is traditional, not documented.
  • The Greek Christian manuscript copied on Arabic paper early in the ninth century is reported; I could not identify it.
  • The dates for Kairouan, Fès and Xàtiva are given as they generally circulate, and were not traced individually.
  • Nāṣir-i Khusraw’s description of goods wrapped in paper in Cairo is often quoted. I could not verify the passage in a reliable edition of the Safarnāma.

Sources

  1. Jonathan M. Bloom, Paper Before Print: The History and Impact of Paper in the Islamic World, New Haven, Yale University Press, 2001, 320 pp.
  2. Adam Benkato (University of California, Berkeley), “Mapping the Mount Mugh Letters and early 8th-century Sogdiana”, Invisible East, University of Oxford, 15 March 2022. Citing V. A. Livshits, Corpus Inscriptionum Iranicarum, Part II, Vol. III, SOAS, 2015, and Sh. Begmatov, Kyoto University, 2020.
  3. Rudolf Sellheim and Mohsen Zakeri, “Ebn al-Nadīm”, Encyclopaedia Iranica, IX/5, pp. 475–483, 15 December 1999, updated 24 January 2012.
  4. Leiden University Library, Legatum Warnerianum Or. 298, Gharīb al-ḥadīth of Abū ʿUbayd al-Qāsim b. Sallām, Dhū al-Qaʿda 252.
  5. “The Second Oldest Dated Complete Book in Arabic on Arabic Paper”, History of Information, citing Bloom 2001, pp. 59–60.

LAYL WA NAHĀR ÉDITIONS

Islamic Art of Living is the editorial home of Layl wa Nahār Éditions. Our first book is in preparation.

Visit the Publishing House and join the list

Comments

What do you think?

Leave a Reply
.
RELATED
Who buys halal: typographic cover reading "The biggest halal export is not food."

Who buys halal?

Everyone cites the size of the halal economy. Almost no one can break it down. Malaysia can, and what its figures show is not what anyone expects.