![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
glaurungHere at last is the third, very long, final part of my three part takedown of Incunabula's viral thread from early September.
To recap: there was a thread on Twitter by Incunabula that went viral last month. In James Burke-ian Connections style, Incunabula said that "Cheese ?? is one of the 5 things the Western book as we know it depends on. The other four are snails, Jesus, underwear and spectacles." The history in that thread ranges from over simplified and inaccurate, through simply wrong, to a disingenuous lie. Today I'm mostly covering the part about "snails" aka Phoenicians, alphabets, and how printing would never have taken off if Gutenberg had used a syllabary or a logosyllabic script instead of an alphabet. Which is built on a lie.
Part one: cheese and Christianity/mostly over simplified and inaccurate. Part two: underwear/simply wrong. The original viral thread that I am laboriously beating to death.
Before tackling snails, let me briefly look at the bit about spectacles. The claim is that that early European printers would not have been able to make money selling books without customers who were able to read into their 40's thanks to spectacles. To this, I will simply point and wave energetically at the history of handwritten paper books in the medieval Islamic world, where individual scholars had five to ten times as many books as the largest monastery libraries in medieval Europe, privately endowed libraries like the House of Wisdom founded by Sabur ibn Ardashir had five times as many books as the largest libraries in all of Europe, and imperial libraries of the Caliphate were 10, 100, or 1,000 times as large as the (anti)Papal library at Avignon, depending on the Caliph and on whether you decide the claimed numbers were exaggerated (see part 2 of this series, linked above).
All of that in a world without printing or spectacles. Spectacles certainly helped book sales, but they were clearly not essential to creating a market for books, or the pre-spectacle Islamic world would not have had such an explosion of books after the introduction of paper.
As for the alphabet's relation to the success of printing. First, I want to get the disingenuous lie out of the way. In the postscript to their thread, Incunabula writes "A characteristic of all syllabic writing systems is that they require far more graphemes - ie symbols - than does an alphabet. Where this is combined with logograms as well, this number can be very high - for example, potentially 50,000+ for Chinese and Japanese."
Galagina Oowatie, publisher of the Cherokee Phoenix newspaper in the early 1800's, would have a few choice words to give Incunabula. The Cherokee Phoenix was a bilingual newspaper, printed with movable type. It carried stories in both English and Cherokee. The Cherokee language portions were printed in the 86 character syllabary invented by Sequoyah.
Some languages (like English) have *lots* of syllables (thousands for English) and are a poor fit for a syllabic writing system. But most syllabaries use less than 200 characters. In an alternative world without the alphabet, relatively simple ancient syllabic systems like Linear B (87 syllable signs plus 100 ideographs) or Cypriot (56 syllable signs) might have taken over Eurasia instead.
China invented movable type centuries before Gutenberg, but printers there only used it for a tenth of the books they printed - usually the largest, most complex printing projects done for an imperial commission. For ordinary print jobs, they used woodblocks, because wrangling tens of thousands of different pieces of type was too difficult.
So that much of what Incunabula says is true. But Chinese, and the writing systems derived from it (Japanese, early Korean, pre-colonial Vietnamese), are unique in that regard. For every other non-alphabetic writing system humans have used, far fewer signs were involved, and printing books with movable type would have been more complex, but not impossibly daunting. Incunabula's waving at Chinese in order to dismiss the viability of all non-alphabetic scripts is dishonest.
Sources for the following: various google search hits, omniglot.com and wikipedia entries on the various writing systems and languages, plus:
1, "Levels of Literacy" by Nick Veldhuis, an article in The Oxford Handbook of Cuneiform Culture, freely available.
2, John Baines' article "Literacy and Ancient Egyptian Society", paywalled but viewable with a free accoun.
3, Jim Loy's Egyptology site sadly defunct but lives on at the Internet Archive.
4, The Digital Egypt for Universities site
5, a book, Brotherhood of Kings by Amanda Podany.
6, my memory of another book I used to own on the decipherment of Egyptian, Linear B, and cuneiform. It was 9x12, with illustrations on glossy paper. The cover was grey. And I cannot remember the title or author. 🙄
All of the oldest writing systems in the world seem to have started out as logographic - signs stood for words. Which gets complicated fast as you start wanting to record more complicated information. Scribes in China said "this is fine" and kept inventing signs until they had thousands of single signs and tens of thousands of compound ones. Everywhere else, the people writing things said, "this is getting cumbersome," and devised a system of signs for isolated syllables or sounds (usually based on rebuses), enabling them to write any word by combining signs. Tradition, the desire for a shorter way to write common words, and the need to disambiguate homophones (especially if the sound symbols lacked some of the sounds used in the language, as with Egypt's vowel-less sound signs) usually kept them from dispensing with logograms altogether.
It's been difficult to impossible for me to determine how complicated many early writing systems were in practice. Modern references are written by and for scholars who are wanting to be able to read all of the texts that exist in the language, over hundreds or thousands of years of the writing system being used. Besides massive gulfs between the oldest and newest texts, there was another gulf between the complete writing system used by trained scribes and that used by ordinary people (once literacy expanded beyond the trained elite, which took centuries if it ever happened at all). Modern scholars, needing to be able to read all kinds of preserved writing from all ages, are faced with learning a significantly more complex writing system than that which individuals at any one time in the past were using to jot notes or write letters.
Looking at writing systems outside East Asia which spread beyond a scribal elite, the full corpus of signs never exceeded a thousand, and was often far fewer. Mayan (the oldest American writing system that can be read today) used about 800 characters: a 150 character syllabary, 100 characters for the names of specific places and gods, and another 550-ish logograms. Middle Egyptian used 700: 24 signs to indicate single consonants, another 180-ish to indicate multiple consonants, and 500-ish determinatives to disambiguate between words that had the same consonants and to mark the end of each word (with around 100 or so determinatives being common and the rest more rare). Any sign in Egyptian could also be used as a logogram by adding a vertical line next to it, but this was usually only done to abbreviate the most common words. Cuneiform was adapted to be used with multiple languages over its three thousand year history. Sumerian and Akkadian (the language that evolved into Babylonian and Assyrian) both used around 600 signs: a syllabary of 120 or so, a couple dozen determinatives (used only for nouns rather than every word), and the rest logograms. Other languages that adapted the cuneiform system to their needs used significantly fewer signs - for example, 375 for Hittite and 130 for Elamite.
In terms of the number of signs needed for day-to-day use, I could not find information on Mayan, Egyptian or Sumerian, but Veldhuis cites a study which found that 170 signs were all that a fully literate Akkadian speaker needed for day to day writing, and says that judging from 19th century BCE letters written by Assyrian merchants in Anatolia, it was possible to get by with even fewer.
For context, a compositor's workstation for setting English type by hand typically consists of a large cabinet of drawers. The top of the cabinet holds two identical trays with about 70 or so compartments in each, total 140-ish compartments (each tray has one style or weight of a-z, A-Z, 0-9, and punctuation). The drawers below the top contain identical trays with less frequently used sizes and weights of type. Ancient languages didn't have case or punctuation, so an imaginary compositor's station for a single style of ancient Akkadian would need to hold about four times as many compartments to be complete, or only a few dozen more if they weren't being asked to set type using unusual signs. If the compositor had three cabinets instead of one (so they stood within a workstation instead of next to one), and each cabinet had an extra 60 compartments, even the completist version of Akkadian could be typeset without too much difficulty.
All of this is moot, of course: outside of East Asia, all pre-alphabetic writing systems died out in common use during the last millennium BCE, replaced by alphabets. Chinese, and the systems derived from it, are the only survivors of the dozens of known pre-alphabetic writing systems (most of which, like Incan quipu, Linear A, the Indus Valley civilization's writing, and the inscriptions found on Easter Island, cannot be read any longer at all).
Why didn't hieroglyphs, cuneiform, or one of the others, take over the world instead? The answer is not, as many sites implicitly assume, that they were too complicated for people to learn. That's just Eurocentric, ABC-centric nonsense: children in China and Japan learn to read and write at a young age, and complexity has far less to do with literacy than the opportunity to study and practice. The far simpler systems used outside of China did not present a stumbling block.
To get to an answer, I'm going to talk a bit about the two ancient writing systems whose histories and fates are best known: cuneiform and Egyptian. Cuneiform's history is quite well understood, because of its writing medium: wet clay. Clay tablets become permanent bricks when heated in a fire. Most cuneiform tablets were never fired and ended up melting back into mud when discarded, but some were fired deliberately, to preserve them for travel or because they were important, and a great many more were fired when the buildings in which they were stored burned - whether accidentally, or as part of the destruction of the old regime by conquering armies.
Starting 10,000 years ago, pre-literate Mesopotamia used fired clay tokens for keeping track of the supply of stored food and other goods. The system evolved over time. After the development of social inequality, tokens representing a debt (goods to be paid as taxes or to a merchant) were sealed inside a clay container: the seal ensured that there was no alteration of the ledger inside, and it could be checked by breaking open the container and tallying the contents. Soon accountants were impressing the tokens onto the outside of the clay container while it was still wet, creating a visible tally of the contents that could be checked without breaking the seal. Two more innovations (and I am not clear which came first) created the first cuneiform writing, which emerged around 3500 BCE: eventually the impressions became the record, obviating the need for tokens or containers in most situations; and also eventually, impressing actual tokens into the clay evolved into inscribing symbols for those tokens into the clay with a stick, eliminating the need for tokens and speeding up the record-making process.
The earliest Sumerian cuneiform consisted of about 1500 signs, and was only used for accounting: it took centuries before the first tablets containing sentences appear. But by 2900-2800 BCE, a simplified and standardized system of 600 signs was being used to write government prose - royal letters and proclamations, legal contracts, and so forth. Literary works eventually also appeared, but written language continued to be largely a government-church monopoly for most of the 3rd millennium. In a world that had no previous concept of writing, the ability of a royal official or priest to look at a slab of clay and know exactly how much wheat you had paid in tribute last year must have seemed magical, and the scribes who knew how to read and write were very conscious of how their work, and the writing they produced, was an expression of the power and majesty of the throne.
And that sense that their work was a tool of great power shaped the system of training scribes. Many tablets containing school assignments for scribes-in-training have survived, such as word lists and rote exercises designed to teach the writing system. And the striking thing is that this scribal curriculum seems to have survived more or less intact for thousands of years. A school from 1800-ish BCE excavated in Nippur provides the most complete view of scribal tutelage, but the bits that have survived from much earlier and much later eras are essentially the same material. Nippur was from the Old Babylonian period and the language the students spoke was Akkadian, but they were made to copy and memorize the same word lists as Sumerian students a thousand years prior. They were expected not to just learn to read and write Akkadian, but to achieve complete mastery of the writing system, in the ancient form of their own language and in the then-dead language of Sumer. The end of the curriculum for Nippur students involved reading Akkadian and Sumerian literature written many centuries before. Especially as Sumerian ceased to be a living language around the end of the third millennium, Cuneiform scribal school became the first classics department.
The Akkadians lived to the north of Sumeria and spoke a completely different language unrelated to Sumerian, but they knew a good idea when they saw it and, in the course of conquering Sumer, they adapted cuneiform for their needs around 2400. In this world, rulers seldom learned to read: when they dictated a letter, they would speak of hearing their correspondent's letter read to them. Especially as the region became the home of a series of large empires, rulers in the provinces or on the periphery were unlikely to even speak Akkadian - but written Akkadian had become the lingua franca for diplomacy and commerce. So scribes were important and fairly influential people, the only ones who could read and translate missives from the empire, the ones charged with turning their employers words into the educated language of diplomacy and government and writing them down in letters, edicts, proclamations, and so on.
In some places, literacy never seems to have moved beyond the circle of trained scribes. In Greece, all of the existing examples of Linear B are written by just a few dozen hands - when the palaces in which the scribes worked fell into ruin during the late bronze age collapse, the writing system became completely forgotten. But in Babylon, writing did eventually spread beyond the community of elite bureaucrats.
Literacy rates in ancient times is a vexed question - the vast majority of people never learned to read, but that's the case in every era up until the early modern period's expectation that all children should attend school before being put to work. But there is a huge difference between a literacy rate of much less than one percent - as for Linear B writers in bronze age Greece - and and literacy of a few percent - enough for there to be a tradition of vernacular writing, enough for the preservation of writing technology to be immune to changes or collapses of government.
Most of the archaeology that's been done in Mesopotamia, especially in the 19th and early 20th centuries, has focused on palaces and the homes of the elite. But some ordinary urban homes have been excavated, and Veldhuis cites a survey of the clay tablets found in those non-elite homes which found a fair number of tablets in the homes of ordinary city dwellers in ancient Babylonian cities, showing that by the second millennium, literacy had spread beyond graduates of the scribal schools to urban people of relatively modest means.
In Egypt, the story is more complicated.
The start of Egyptian writing is still murky because Egyptians lived on the flood plain of the Nile - the remains of most ancient towns are deeply buried under alluvial sediments and unfindable by archaeologists. Plus the damp soil by the river works against preservation of artefacts other than stone and pottery. Away from the river was desert where very few people lived - but people went there to worship and bury their dead. It's easy to find and excavate things in the desert, and perishable items like papyrus have survived well there, but with a tiny list of exceptions, you're not digging up places where people actually lived (exceptions include towns where quarry workers and tomb builders lived, and a few waypoints along roads built through the desert to connect cities on the Nile where the river was too curved to be the fastest route).
Aside from those exceptions, which have yielded graffiti and day to day writings on papyrus and ostraka (pot shards used for note taking), our knowledge of writing in Egypt is filtered through the items that were buried with dead people or inscribed on temples/monuments. I can't count the number of times I've seen someone say that ancient Egyptian culture was obsessed with death and the afterlife, and, no, that is a mistaken impression based on the overwhelming majority of the writing we've found being funerary.
The earliest preserved writing yet found, from the 4th millennium, appears as brief labels on grave goods - something like warehouse inventory tags. A bit later, murals in tombs start to have captions (although some tombs have a place for a caption but nothing written there - writing was ancillary and supplemental in tombs for a long time. The Great Pyramid, famously, contains no murals or inscriptions at all). The limited evidence found suggests a gradual evolution similar to cuneiform, from a writing system used mainly as an administrative tool in the 4th millennium to one that could record literature, letters, and biographies in the mid 3rd millennium. While pre-literate Sumer and Egypt were aware of each other and traded, there's zero evidence of technology transfer WRT nascent writing systems (scholars continue to argue over whether Egyptian writing was inspired by Sumerian writing, but as new finds continue to push the date of the earliest known hieroglyphs further back, it looks more and more like the amount of inspiration was extremely minimal to non-existent).
Very quickly after Egyptian writing moved beyond the warehouse and account books, it became a marker of prestige and power. Elites in Egypt often included the fact that they were a scribe in their biographies or official titles. Not that they spent a lot of time doing scribal work - that was relegated to clerks - but that they had the skill. The *fact* that they could read and write enhanced their power and prestige. So literacy became near-ubiquitous among rulers, priests, and the rest of the 1%.
Elites saying "this is our thing" helped give writing a very different path than in Mesopotamia. We have some glimpses into literacy amongst ordinary people - such as papyrus and ostraka found in Deir el-Medina, the village of excavators and artisans who built tombs in the Valley of Kings from 1500-1000. But the overwhelming majority of preserved writing was produced by or for the ruling class, and they liked keeping their toy to themselves. One tool of exclusion was the way the writing system did not evolve to keep pace with changes in the spoken language.
From 2500 to 500-ish, you had a very stable writing system - but in that time the spoken language transformed radically. Egyptian language evolved from synthetic (prefixes and suffixes indicate grammar) to analytic (helper words indicate grammar), so the whole structure of the language had changed. Plus, of course, evolution in pronunciations meant that many of the rebuses built into the sign system no longer worked as clues to how to interpret the signs. Imagine if people spoke English, but it was impossible to write anything except in classical Latin? Like that. Numerous examples of monumental inscriptions have been found where the scribe slipped up and produced a text that was more like how they spoke than how they were supposed to write.
Technical side note: Egyptian writing had two forms, hieroglyphs which were used for inscriptions on stone in tombs and on monuments, and hieratic writing, which was a simplified version of the hieroglyph signs optimized for handwriting on papyrus. These started out as merely different fonts of the same writing system in the same language.
Around 1350 BCE, people finally cracked under the strain of not having a writing system for the language they spoke, and in hieratic documents, especially literary works, you get Late Egyptian, with the same sign system but using analytic grammar, changes in the consonantal values used for writing words, and generally shifts which (scholars assume) moved the writing system closer to how people were talking. At the same time, in hieroglyphic inscriptions, Late Egyptian signifies a different development, with a huge explosion of new signs and much more complicated texts.
Another side note: Hieroglyphs had always been very pictorial, consisting of easily recognized little drawings of people, animals, plants, etc. To recap, each sign had a sound meaning (often but not always based on a rebus of the word for the thing depicted) or a determinative meaning (again not always related to the thing depicted), but it could also have a logographic meaning (yet again not always clearly connected to the depiction). In short, the system lent itself to cryptic writing - deliberately choosing obscure signs, seldom used logograms, complicated consonantal signs instead of common ones, and so forth, to produce texts that took extra effort to read. Taking it a step further, you could easily create complications to the system to write in code, using special rules that made your text unreadable to someone not in the know.
Late Egyptian hieroglyphic inscriptions doubled down on the cryptic potential of hieroglyphs, adding thousands of new signs in the process and generally going out of its way to be harder to read - and while no sources I found talk much about *why* this was done, it might have been a reactionary response to increased literacy among non-elites, a desire to keep the system "ours" and not let the hoi polloi have it.
And then there was another upheaval around 500 BCE, and hieratic writing ceased to be used except in religious contexts. Instead an even more simplified handwritten sign system called demotic emerged, which was used for everything else. Demotic altered the sign system substantially, again bringing it closer to how people spoke (which had evolved since the 1300s) as well as simplifying the individual signs to be even easier to write. Scholars today can connect demotic signs to their pictorial forebears, but looking at them in isolation, they were no longer recognizable little pictures of things - which helped since those pictures hadn't been a useful guide to meaning for a long time. I haven't found any explicit discussion of the number of signs needed to write demotic, but a presentation by Irena Ristevska on youtube includes a slide saying her team needed about 500 glyphs to create a digital font of the demotic script on the Rosetta Stone.
Returning to the question of why neither cuneiform nor hieroglyphs/hieratic took over the world the way alphabets did:
On the one hand, both scripts were at least somewhat burdened by complexities of adaptation that alphabets did not suffer from nearly as much. Syllabaries do not travel well - you have to reinvent the syllabary for each new language you port the writing system to. Isolated consonants and vowels vary a lot less between languages. As for Egyptian's system, the rebuses and the pictorial nature of the script became a hindrance, not a help, to someone trying to adapt the system to a new language. Egypt had the 24 single consonant signs from the beginning, but no one ever took those as the basis for a simplified, stripped down system the way the Hittites and Elamites did with cuneiform.
On the other hand, the biggest factor seems to have been the contingencies of history. A long-lived, incredibly stable scribal education system and culture in Mesopotamia created a pool of scribes who would happily translate Akkadian writing into languages understood by their employers, vastly reducing the need to adapt the writing system to the local language in lands on the periphery of Mesopotamia. Which largely cut off demand for adapted, localized writing systems in peripheral kingdoms - and that lack of adaptation made it hard for the technology to spread out beyond the sphere of Akkadian/Babylonian/Assyrian influence to lands outside the ken of Mesopotamia's scribal schools. The Hittites seem to be the farthest flung nation that used cuneiform.
Geography was against Egypt's writing system spreading far beyond the Nile - with deserts to the east, west, and northeast, the only close neighbours Egypt had were on the upper reaches of the Nile or its tributaries. Kush (modern Sudan) began using hieroglyphs around 1000 BCE, after they conquered Egypt for a time. If ancient Ethiopia started using hieroglyphs, I've found no reference to it. Other than those two, Egypt's closest contacts were in the Mesopotamian sphere and already had a perfectly good writing system, TYVM. But most of all, Egypt's writing system was "owned" by elites who had an interest in keeping it complicated and exclusive.
As for alphabets:
Sometime before the end of the 19th century BCE, a group of non-Egyptians who spoke an early Semitic language, and who lived in or next to Egypt, devised a consonantal alphabet (an abjad) based on 25 or so Egyptian hieratic signs. The earliest examples of the script in question have been found in two places, one as graffiti at Wadi el-Hol, on the desert road between two Egyptian cities, dated to the 19th century BCE; and the other in a turquoise mine and in the temple beside the mine, in Sinai, which is known to have been actively worked in the 19th and 16th centuries. Called Proto-Sinaitic after the place where it was first found, the script clearly does not use Egyptian sound values for the signs. However, if you assume the script is for a Semitic precursor to Canaanite/Phoenician/early Hebrew (originally all the same language), and use those sound values, it's possible to read a few words. Thousands of very short, often fragmentary examples of Proto-Sinaitic have been found throughout the Palestine-Lebanon region, with various 2nd millennium dates. Nobody knows anything about who invented Proto-Sinaitic, and the extremely short, fragmentary finds make it unclear how and by whom it was being used over the course of the 2nd millennium, but it was clearly being used in the region and a line can be drawn from it to the emergence of written Phoenician around 1000 BCE (keeping in mind that while Phoenicia the maritime powerhouse was a specific group of city states in what is now Jordan, Phoenician culture, language and writing were mostly indistinguishable from those of the people living in Canaan and Palestine at the time).
During the 2nd millennium, the city-states that later became Phoenicia were important to Egypt only as a source of cedar wood and a link on the trading roads to Mesopotamia. Egypt conquered them but allowed them autonomy in exchange for wood and free passage to the east. As maritime powers go, the city states were nothing compared to the Mycenaean civilization in Greece and Crete. Then, in 1200, a poorly understood series of calamities dubbed the Late Bronze Age Collapse destroyed the Mycenaean civilization and badly weakened Egypt - and the Phoenicians stepped into the resulting power vacuum. They built ships and created a powerful trans-Mediterranean trading network, founding or conquering numerous North African port cities, and becoming dominant sea traders in the region until the re-emergence of Greece in the 9th century (note that they were never politically unified, so there was never a Phoenician empire, just several cities each with its own trading network). And after 1000, everywhere they went, Phoenicians took their alphabet with them. Would they have spread so far if they hadn't been the only source of Tyrian purple dye, made from those snails Incunabula talks about? Probably. The dye certainly helped, but fundamentally they were seafaring traders, and if we must point to a single emblematic regional product that powered their rise, it would be the cedars that they used to build their ships.
But that alphabet of theirs didn't just travel west - it also went east, and neither snails nor cedar had anything to do with that spread. The Ancient Aramaic alphabet emerged in the Persian empire around 800 BCE. Because Persia was large and had many languages, there was also the Old Persian cuneiform writing system - starting around 500, Persian speakers junked the syllabary/determinative/logogram edifice of the past three thousand years of cuneiform writing and instead started to use a small set of cuneiform signs as an alphabet. Meanwhile, Proto-Sinaitic developed into Ancient South Arabian also around 800, which in turn inspired alphabets in Ethiopia and the Horn of Africa.
In southern India, the Brahmi alphabetic script appears suddenly around the 6th-4th century BCE, with no easily traceable path back to alphabets from the west (Persia never spread further than Pakistan and the Indus valley; India traded with Arabia but there's no recognizable connection between the either Persian or West African/South Arabian scripts and Brahmi). And yet, there's no identifiable predecessor to Brahmi in India either (between the extinction of the undeciphered Indus valley writing system in 1900 and the emergence of Brahmi ~1500 years later, no sign of writing existing in India has been found). The question of whether the Indian alphabets were independently invented or owe their genesis to scripts from the west remains a vexed question for scholars, made more contentious by Indian nationalists, including some scholars, who very strongly *want* India to have invented its own alphabet from scratch without outside influence or inspiration. Whatever Brahmi's origins, writing systems based on it spread throughout India to Tibet, Southeast Asia, and eventually to Korea, where Sejong the Great, inspired by the Tibetan script, invented the Hangul alphabet in 1443 CE as a supplement or replacement for Chinese characters, which had been used in Korea up until that time but were ill-suited to spoken Korean (Chinese is analytic; Korean is an agglutinative synthetic language).
Sejong is thought to have been inspired by Tibetan writing, but he invented a unique system designed to be simple and fast for the common people to learn. The populist slant of Hangul was frowned upon by Korean nobility, and in 1504 Hangul was banned after the publication of a tract criticizing King Yeonsangun, but the ban did not last and with minimal modification Hangul continues to be used in Korea today.
Which brings us, finally, to printing and movable type. (sources: wikipedia, the Chinese sections of Jonathan Bloom's "Paper Before Print," some Reddit threads, google searches based on memories of a book I read in college 33 years ago, and a paywalled article, "Early Korean Printing" by Pow-key Sohnn.
Woodblock printing originated in China sometime before 220 CE and for a long time was mainly used to print pretty patterns on silk. Buddhists seem to have been the first to realize that you could use woodblocks to print many copies of writing on paper. They were not interested, at first, in doing this to to make books to be read. Mahayanan Buddhism regards religious texts, as physical objects, to have talismanic power, and the creation of such texts as a meritorious act. Woodblock printing made it possible to produce hundreds or thousands of textual talismans, and thus generate far more merit than was possible with hand copying a text. Buddhists began to use printing to churn out such not-intended-for-reading texts in the 600's. Some were distributed, but many more were simply ritually buried en masse after being made - the point was the making of them. But printed Buddhist texts that were meant to be read started to appear very soon afterwards - the earliest known is thought, based on analysis of the character forms, to have been printed at the end of the 600's. Non-Buddhist books began to be printed in significant numbers in the 800's, and printing started to become quite common in the 900's.
Woodblock printing spread quickly to Korea and Japan - In Korea the earliest printed book found dates to the early 700's; in Japan around 770 Empress Shotoku commissioned the printing of a million copies of a Darani sutra for talismanic purposes, with each copy encased in a miniature pagoda. Printing of books meant to be read was happening by the 900's in Korea and the 1000's in Japan.
Histories of printing almost always are actually histories of printed books, and neglect how printing transformed realms other than books. For a government bureaucracy, printing is obviously extraordinarily useful, allowing mass production of flyers, posters, forms, and so on. But I couldn't find anything online about when exactly the imperial government embraced printing, except with reference to imperial commissioned editions of books.
In the late 600's emperor Wu Zetian expanded and made permanent the nascent imperial examinations system. Aspiring examinees needed copies of the canonical Thirteen Classics of Confucian tradition to study, but seem to have had to make do with handwritten copies until the 930's, when the government commissioned a printed edition of those classics. Closer to the end of the 900's, the first few Song emperors supercharged the uptake of printing with a vast expansion of the examinations system. A text from 1005 describes a visit by the emperor to the imperial Publications Office, which had 100,000 woodblocks in its warehouse at the time, in contrast to the mere 4,000 blocks that had been in stock at the start of the dynasty's regime in 960. The publications officer went on to enthuse about how, thanks to printing, it was so much easier to obtain copies of the classics under the rule of the current emperor and his predecessor than it had been under the previous dynasty.
Handwritten books continued to be more prestigious than printed ones. Book collectors prized and desired handwritten books above printed ones, but printed books were *ten times* cheaper - which means they became far more abundant. In all of Chinese history before 960, a total of 200 private libraries of books are known, and only 140 of those are from the era before printing began to take off. In the next four centuries (the Song dynasty), we know of 700 private collections. The sizes of those collections increased dramatically after printing as well. Chinese retained the term "juan" (scroll) as a unit of book measurement even after the rise of codexes, which conveniently allows direct comparison between scroll based (pre 900's) and codex based (post 1000) libraries. The pre-printing library of 5th century scholar Shen Yue had 20,000 juan. In the 1100's, the scholar Ye Mengde had 100,000 juan of books totalling 6,000 titles.
The huge increase in the availability of books boosted literacy. In 1488, a ship of Koreans, including a government official Choe Bu, was blown far off course by a multi-day storm and found themselves on the coast of Southern China near Ningbo. As a government official, Choe Bu could read and write Chinese characters, and he found that despite not knowing a word of spoken Chinese, during the shipwrecked Koreans' long trip back home, he could communicate in writing with almost everyone he met - "even village children, ferrymen and sailors."
Woodblock printing technology started with writing the text normally on translucent paper, which was pasted upside down onto a wooden block. Then a wood carver would chisel away a layer of wood everywhere that the paper had no ink, leaving only a mirror image of the desired characters. If a mistake escaped the initial proofreading stages and got carved into the block, it could be chiselled out and replaced with an inlay containing the correct text. The finished block would be inked and a sheet of paper pressed firmly onto it. Since the block was not being smashed down onto the paper as is done with European-style printing, the paper could be relatively thin and delicate. The most time consuming part was obviously carving the wooden blocks; for big projects, storing and wrangling all the blocks (in case more copies were needed later) could get challenging - in Korea, a 971 edition of the Bhuddist Tripitaka scriptures used 130,000 blocks.
In 1040, a Chinese printer named Bi Sheng decided he was tired of dealing with so many blocks and invented movable type. He experimented with making the individual pieces of type from wood, but quickly discarded that in favour of ceramic. The scholar Shen Kuo wrote a detailed account of his invention and process, talking about how most words had "several" pieces of type for them while more common words had "twenty or more"; how all of these many thousands of pieces of type were organized by rhyme group with paper labels; and how Bi Sheng usually had one page being printed while he composited the next; and how "if one were to print only two or three copies, this method would be neither simple nor easy. But for printing hundreds or thousands of copies, it was marvellously quick."
Ceramic pieces of type had to be individually sculpted, and sometimes slumped or warped in the kiln. Making type from bronze, cast from moulds using wooden masters, produced much more uniform results. Paper money started being printed with bronze type in the 1100's. In 1190, Zhou Bida read Shen Kuo's account and made ceramic type for printing his book. In 1234, the first books made with metal type were printed in Korea. Around 1300, in China, Wang Zhen published a book on how to print books using wooden type. He also wrote about his system for sorting the type: he used a pair of rotating round (multi level?) tray tables on which compartments of type were sorted by tone and rhyme; the compositor could spin the tables to bring the next needed word into easy reach.
Wood, metal and sometimes ceramic type continued to be used to print some books in China, but woodblock printing remained the default for 90% of printing until the introduction of European mechanised printing methods. There were a few obvious reasons why movable type did not displace woodblocks in China. First, upfront costs: woodblock printing required wood and a wood carver, who did not need to be literate. Movable type required the creation of tens of thousands of pieces of type, minimum, before you could print a single sheet. Second, training: any carpenter could carve woodblocks. Movable type required setting up a complex organizational scheme for storing the type, and then a highly literate compositor had to learn that scheme by heart. Third, scale: setting movable type was more economical only for large jobs. For the small jobs which are the bread and butter of any printing business, it made more sense to just make woodblocks.
Things were a bit different in Korea, which did not have many hardwood trees suitable for making woodblocks: metal type became much more common in Korea for this reason, but printing remained a government monopoly, with independent printers and the selling of books strongly discouraged. So there wasn't the explosion of books and literacy that occurred in China. That government monopoly also prevented movable type technology from spreading to Japan until the late 1500's, by which time Jesuits from Europe had already arrived with printing presses of their own (Japan's relations with China were at arm's length due, among other things, to a long term problem with Japanese pirates preying on Chinese ships).
Incunabula claims that printers in East Asia experimented with movable type but gave up on it. Which is a common claim in Eurocentric histories, but it's not true. Movable type became a known technology, and a steady trickle of books were produced with it in China - but only when it made financial sense to do so. For most jobs, making woodblocks was cheaper. It was only for large jobs, mainly government commissions, that movable type made sense. Behind that steady trickle of movable type books was a small group of printers who made the heavy investment in type, memorized a storage system, and sold their services to the government. Movable type would have had a larger share of the printing market if China had a tradition of large print runs, but for whatever reason, small runs of a hundred or so copies were much more typical.
Printing and the Korean alphabet: In June of this year, a trove of bronze type was dug up in Seoul that included some of the earliest Hangul type yet found, from around 1460 - but the cache contained more Chinese type than Hangul type. The Hangul alphabet and movable type didn't get a chance to show their synergistic power in Korea for a long time, mainly because the 1% in Korea at the time were Sinophiles, so their desire was for printed editions of Chinese literature. For writing in Korean, the elites preferred to use Chinese characters supplemented by Hangul for the grammatical affixes and particles that did not exist in Chinese. The lower classes and women, who were the primary users of pure Hangul, were not able, or not allowed, to set up printing operations.
Printing also moved west. India started making cotton fabric prints in the 900's. Never having adopted paper, they did not start printing books there. In the Islamic world, block printing was mainly used for making charms and talismans; the earliest known, from Egypt, also dates to the 900's. Arabic script was entirely cursive, and movable type is really only suited to scripts with separation between the letters (the first Arabic type fonts that produced acceptable looking results were not made until the modern era). The same feeling that handwritten books were superior existed in the Islamic world as in China, only much stronger. Completely aside from the issues with making attractive Arabic type, calligraphy in the Islamic world was a highly prized art form, and there was zero interest there in producing cheaper yet less beautiful books by using woodblock printing.
Because of this, printing arrived in Christian Europe as a technology for printing patterns on fabrics, completely divorced from any idea that it was useful for making books. Paper arrived a little bit after printing, and by the 1400's, printing started being used with paper to produce art prints, playing cards, and such. The earliest wood block printed European books were heavily illustrated, and were being made by 1440 if not earlier. From that as a starting point, either Gutenberg had the same epiphany as Bi Sheng five hundred years before, or someone who had been to China told him about movable type. Nobody knows which it is - Gutenberg wrote nothing about his invention, himself, or his printing shop, and most of what is known about the press he made is inferred from copies of his Bible and from other early printers who copied his setup, and did write about their tools and procedures. He does seem to have been the first person to reverse the relation of the type/block and the paper, using a screw to press the inked type down onto the paper rather than pressing the paper onto the inked surface.
In Europe, the movable type printing press brought about a massive series of cultural changes which ended the medieval period, splintered a monolithic Christianity into competing sects, enabled a rise in scholarship and powered a series of revolutions in science. There's a tendency among European historians to say that it was the synergistic combination of an alphabet and movable type that made all that possible, but looking at China, I don't think that's right. And there's another trend to say that China and Korea had strong unified governments that could tell printers and scholars what to do and what to study, whereas 15th century Europe had tiny fragmented polities ruled by incredibly weak kings, and that enabled more intellectual freedom, less censorship, and thus more technological and cultural experimentation and change. But I don't think that's it either.
Europe in the 1400's was incredibly poor and powerless, but when printing was still just beginning there, Europeans were already working to enrich themselves, and were already travelling down the path that led them to rapacious conquest, enslavement, colonization, and general global bullying. They had already started to have a period of intellectual ferment and technological change based on rediscovery of lost classics and new innovations, and that ferment was being fuelled in part by greed and the desire to devise better tools for warfare. Printing just pushed things along faster.
When post-Gutenberg Europeans first arrived in China and looked around, they (mostly very bigoted Jesuit missionaries) saw a culture where books were not just ubiquitous, but actually as cheap or cheaper than they were in Europe. They spoke of that as a bad thing (because the books were Confucian or Buddhist, and they were Christian bigots). But my takeaway is that China didn't need movable type to have an explosion of literacy, of books, and of scholarship. It had already had its own printing revolution, about six hundred years before Europe's. But that revolution expressed itself in a very different form, with very different long term results. China was already rich, already powerful. Instead of a quest for wealth and power that in turn fuelled fast scientific and technological developments, printing in China powered a quest to keep corruption in check and the exploitation of the poor within acceptable limits, by instituting government by those who had deeply studied Confucian principles.
Looking at the world we've gotten thanks to technology, science, and rampant exploitation fuelled by European greed, I can't say China was foolish to take the path it did.
Bit that didn't fit above:
The Dunhuang Manuscripts are a cache of scrolls and codexes found in 1900 by a Buddhist monk who was looking to rehabilitate a disused monastery in Western China. Totalling about 500 cubic feet of paper, the manuscripts had been sealed inside a cave in the early 1000's. They're an incredibly rich source of historical information, as well as a treasure trove of knowledge about Chinese paper and book making from the 4th to 10th centuries. Sadly, the dysfunctional nature of China's government of the early 20th century meant that when the discovery was reported to authorities, they hemmed and hawed about the expense of warehousing and moving so many books back east for conservation and study. After a few years of getting nowhere, the discoverer of the cave, needing money for his project, sold many of the manuscripts to European looters.
The usual history of Chinese bookmaking posits a neat, orderly succession of formats from scrolls through various refinements of the codex - pasting the pages together at a spine, wrapping the spine, and finally sewing the pages instead of gluing them. Looking at examples from Dunhuang that ended up in the British Library, Li Zhizhong's "Problems in the history of Chinese Bindings" shows how instead of an orderly progression of innovation, all of those formats were experimented with in the 900's before the "butterfly binding" won out in the 1000's.
The codex format spread from Rome to Persia and Arabia but foundered in the steppes and deserts between Persia and China. The earliest paper books in China mimicked the scroll format set in the era of writing on bolts of silk: manuscript pages written on one side were pasted together, then rolled up. All of the pre-900's printed books which survive are scrolls. A scroll got unwieldy with longer manuscripts, and in the 900's book makers seem to have tried various alternative formats which can be found in the Dunhuang cache, and all of which are spoken about in Chinese histories, but as representing a succession of formats. "Whirlwind" or "dragon scale" binding was a compromise adaptation of the scroll format: it pasted individual manuscript pages (written on both sides) to a relatively short blank scroll, which could then be rolled up and stored like a scroll. Unrolled, you had a long strip of overlapping pages, looking rather like scales, which could be turned through. The next step was to paste the pages much closer together onto a much smaller backing paper, producing something halfway in between a whirlwind scroll and a codex - a format found in Dunhuang but not attested to in historical records. The final form that caught on for the next few centuries was "butterfly" binding - which dispensed with the scroll entirely and directly pasted a stack of folded pages together along the fold, producing a codex that fell open like a butterfly's wings. Since Chinese woodblock printing usually used thin paper printed on only one side, you had a pair of blank pages between every pair of printed pages. The butterfly format had staying power and remained very popular from 1000 through the 1300's, when printers started using "wrapped back" binding. This reversed the glue point to the edges of the folded sheet, hiding the blank side of the paper within the fold. Gluing a backing onto the spine helped strengthen the book and guarded against the pages falling apart from glue failure. Finally, "wrapped back" books started being sewn instead of glued in the 1400's... but again, sewn manuscripts were found in the Dunhuang cache. All of those book formats were printed on very thin paper, so Chinese books tended to be quite floppy and delicate. Sometimes they were strengthened with stiff covers; other times each book was stored in its own small box.
To recap: there was a thread on Twitter by Incunabula that went viral last month. In James Burke-ian Connections style, Incunabula said that "Cheese ?? is one of the 5 things the Western book as we know it depends on. The other four are snails, Jesus, underwear and spectacles." The history in that thread ranges from over simplified and inaccurate, through simply wrong, to a disingenuous lie. Today I'm mostly covering the part about "snails" aka Phoenicians, alphabets, and how printing would never have taken off if Gutenberg had used a syllabary or a logosyllabic script instead of an alphabet. Which is built on a lie.
Part one: cheese and Christianity/mostly over simplified and inaccurate. Part two: underwear/simply wrong. The original viral thread that I am laboriously beating to death.
Before tackling snails, let me briefly look at the bit about spectacles. The claim is that that early European printers would not have been able to make money selling books without customers who were able to read into their 40's thanks to spectacles. To this, I will simply point and wave energetically at the history of handwritten paper books in the medieval Islamic world, where individual scholars had five to ten times as many books as the largest monastery libraries in medieval Europe, privately endowed libraries like the House of Wisdom founded by Sabur ibn Ardashir had five times as many books as the largest libraries in all of Europe, and imperial libraries of the Caliphate were 10, 100, or 1,000 times as large as the (anti)Papal library at Avignon, depending on the Caliph and on whether you decide the claimed numbers were exaggerated (see part 2 of this series, linked above).
All of that in a world without printing or spectacles. Spectacles certainly helped book sales, but they were clearly not essential to creating a market for books, or the pre-spectacle Islamic world would not have had such an explosion of books after the introduction of paper.
As for the alphabet's relation to the success of printing. First, I want to get the disingenuous lie out of the way. In the postscript to their thread, Incunabula writes "A characteristic of all syllabic writing systems is that they require far more graphemes - ie symbols - than does an alphabet. Where this is combined with logograms as well, this number can be very high - for example, potentially 50,000+ for Chinese and Japanese."
Galagina Oowatie, publisher of the Cherokee Phoenix newspaper in the early 1800's, would have a few choice words to give Incunabula. The Cherokee Phoenix was a bilingual newspaper, printed with movable type. It carried stories in both English and Cherokee. The Cherokee language portions were printed in the 86 character syllabary invented by Sequoyah.
Some languages (like English) have *lots* of syllables (thousands for English) and are a poor fit for a syllabic writing system. But most syllabaries use less than 200 characters. In an alternative world without the alphabet, relatively simple ancient syllabic systems like Linear B (87 syllable signs plus 100 ideographs) or Cypriot (56 syllable signs) might have taken over Eurasia instead.
China invented movable type centuries before Gutenberg, but printers there only used it for a tenth of the books they printed - usually the largest, most complex printing projects done for an imperial commission. For ordinary print jobs, they used woodblocks, because wrangling tens of thousands of different pieces of type was too difficult.
So that much of what Incunabula says is true. But Chinese, and the writing systems derived from it (Japanese, early Korean, pre-colonial Vietnamese), are unique in that regard. For every other non-alphabetic writing system humans have used, far fewer signs were involved, and printing books with movable type would have been more complex, but not impossibly daunting. Incunabula's waving at Chinese in order to dismiss the viability of all non-alphabetic scripts is dishonest.
Sources for the following: various google search hits, omniglot.com and wikipedia entries on the various writing systems and languages, plus:
1, "Levels of Literacy" by Nick Veldhuis, an article in The Oxford Handbook of Cuneiform Culture, freely available.
2, John Baines' article "Literacy and Ancient Egyptian Society", paywalled but viewable with a free accoun.
3, Jim Loy's Egyptology site sadly defunct but lives on at the Internet Archive.
4, The Digital Egypt for Universities site
5, a book, Brotherhood of Kings by Amanda Podany.
6, my memory of another book I used to own on the decipherment of Egyptian, Linear B, and cuneiform. It was 9x12, with illustrations on glossy paper. The cover was grey. And I cannot remember the title or author. 🙄
All of the oldest writing systems in the world seem to have started out as logographic - signs stood for words. Which gets complicated fast as you start wanting to record more complicated information. Scribes in China said "this is fine" and kept inventing signs until they had thousands of single signs and tens of thousands of compound ones. Everywhere else, the people writing things said, "this is getting cumbersome," and devised a system of signs for isolated syllables or sounds (usually based on rebuses), enabling them to write any word by combining signs. Tradition, the desire for a shorter way to write common words, and the need to disambiguate homophones (especially if the sound symbols lacked some of the sounds used in the language, as with Egypt's vowel-less sound signs) usually kept them from dispensing with logograms altogether.
It's been difficult to impossible for me to determine how complicated many early writing systems were in practice. Modern references are written by and for scholars who are wanting to be able to read all of the texts that exist in the language, over hundreds or thousands of years of the writing system being used. Besides massive gulfs between the oldest and newest texts, there was another gulf between the complete writing system used by trained scribes and that used by ordinary people (once literacy expanded beyond the trained elite, which took centuries if it ever happened at all). Modern scholars, needing to be able to read all kinds of preserved writing from all ages, are faced with learning a significantly more complex writing system than that which individuals at any one time in the past were using to jot notes or write letters.
Looking at writing systems outside East Asia which spread beyond a scribal elite, the full corpus of signs never exceeded a thousand, and was often far fewer. Mayan (the oldest American writing system that can be read today) used about 800 characters: a 150 character syllabary, 100 characters for the names of specific places and gods, and another 550-ish logograms. Middle Egyptian used 700: 24 signs to indicate single consonants, another 180-ish to indicate multiple consonants, and 500-ish determinatives to disambiguate between words that had the same consonants and to mark the end of each word (with around 100 or so determinatives being common and the rest more rare). Any sign in Egyptian could also be used as a logogram by adding a vertical line next to it, but this was usually only done to abbreviate the most common words. Cuneiform was adapted to be used with multiple languages over its three thousand year history. Sumerian and Akkadian (the language that evolved into Babylonian and Assyrian) both used around 600 signs: a syllabary of 120 or so, a couple dozen determinatives (used only for nouns rather than every word), and the rest logograms. Other languages that adapted the cuneiform system to their needs used significantly fewer signs - for example, 375 for Hittite and 130 for Elamite.
In terms of the number of signs needed for day-to-day use, I could not find information on Mayan, Egyptian or Sumerian, but Veldhuis cites a study which found that 170 signs were all that a fully literate Akkadian speaker needed for day to day writing, and says that judging from 19th century BCE letters written by Assyrian merchants in Anatolia, it was possible to get by with even fewer.
For context, a compositor's workstation for setting English type by hand typically consists of a large cabinet of drawers. The top of the cabinet holds two identical trays with about 70 or so compartments in each, total 140-ish compartments (each tray has one style or weight of a-z, A-Z, 0-9, and punctuation). The drawers below the top contain identical trays with less frequently used sizes and weights of type. Ancient languages didn't have case or punctuation, so an imaginary compositor's station for a single style of ancient Akkadian would need to hold about four times as many compartments to be complete, or only a few dozen more if they weren't being asked to set type using unusual signs. If the compositor had three cabinets instead of one (so they stood within a workstation instead of next to one), and each cabinet had an extra 60 compartments, even the completist version of Akkadian could be typeset without too much difficulty.
All of this is moot, of course: outside of East Asia, all pre-alphabetic writing systems died out in common use during the last millennium BCE, replaced by alphabets. Chinese, and the systems derived from it, are the only survivors of the dozens of known pre-alphabetic writing systems (most of which, like Incan quipu, Linear A, the Indus Valley civilization's writing, and the inscriptions found on Easter Island, cannot be read any longer at all).
Why didn't hieroglyphs, cuneiform, or one of the others, take over the world instead? The answer is not, as many sites implicitly assume, that they were too complicated for people to learn. That's just Eurocentric, ABC-centric nonsense: children in China and Japan learn to read and write at a young age, and complexity has far less to do with literacy than the opportunity to study and practice. The far simpler systems used outside of China did not present a stumbling block.
To get to an answer, I'm going to talk a bit about the two ancient writing systems whose histories and fates are best known: cuneiform and Egyptian. Cuneiform's history is quite well understood, because of its writing medium: wet clay. Clay tablets become permanent bricks when heated in a fire. Most cuneiform tablets were never fired and ended up melting back into mud when discarded, but some were fired deliberately, to preserve them for travel or because they were important, and a great many more were fired when the buildings in which they were stored burned - whether accidentally, or as part of the destruction of the old regime by conquering armies.
Starting 10,000 years ago, pre-literate Mesopotamia used fired clay tokens for keeping track of the supply of stored food and other goods. The system evolved over time. After the development of social inequality, tokens representing a debt (goods to be paid as taxes or to a merchant) were sealed inside a clay container: the seal ensured that there was no alteration of the ledger inside, and it could be checked by breaking open the container and tallying the contents. Soon accountants were impressing the tokens onto the outside of the clay container while it was still wet, creating a visible tally of the contents that could be checked without breaking the seal. Two more innovations (and I am not clear which came first) created the first cuneiform writing, which emerged around 3500 BCE: eventually the impressions became the record, obviating the need for tokens or containers in most situations; and also eventually, impressing actual tokens into the clay evolved into inscribing symbols for those tokens into the clay with a stick, eliminating the need for tokens and speeding up the record-making process.
The earliest Sumerian cuneiform consisted of about 1500 signs, and was only used for accounting: it took centuries before the first tablets containing sentences appear. But by 2900-2800 BCE, a simplified and standardized system of 600 signs was being used to write government prose - royal letters and proclamations, legal contracts, and so forth. Literary works eventually also appeared, but written language continued to be largely a government-church monopoly for most of the 3rd millennium. In a world that had no previous concept of writing, the ability of a royal official or priest to look at a slab of clay and know exactly how much wheat you had paid in tribute last year must have seemed magical, and the scribes who knew how to read and write were very conscious of how their work, and the writing they produced, was an expression of the power and majesty of the throne.
And that sense that their work was a tool of great power shaped the system of training scribes. Many tablets containing school assignments for scribes-in-training have survived, such as word lists and rote exercises designed to teach the writing system. And the striking thing is that this scribal curriculum seems to have survived more or less intact for thousands of years. A school from 1800-ish BCE excavated in Nippur provides the most complete view of scribal tutelage, but the bits that have survived from much earlier and much later eras are essentially the same material. Nippur was from the Old Babylonian period and the language the students spoke was Akkadian, but they were made to copy and memorize the same word lists as Sumerian students a thousand years prior. They were expected not to just learn to read and write Akkadian, but to achieve complete mastery of the writing system, in the ancient form of their own language and in the then-dead language of Sumer. The end of the curriculum for Nippur students involved reading Akkadian and Sumerian literature written many centuries before. Especially as Sumerian ceased to be a living language around the end of the third millennium, Cuneiform scribal school became the first classics department.
The Akkadians lived to the north of Sumeria and spoke a completely different language unrelated to Sumerian, but they knew a good idea when they saw it and, in the course of conquering Sumer, they adapted cuneiform for their needs around 2400. In this world, rulers seldom learned to read: when they dictated a letter, they would speak of hearing their correspondent's letter read to them. Especially as the region became the home of a series of large empires, rulers in the provinces or on the periphery were unlikely to even speak Akkadian - but written Akkadian had become the lingua franca for diplomacy and commerce. So scribes were important and fairly influential people, the only ones who could read and translate missives from the empire, the ones charged with turning their employers words into the educated language of diplomacy and government and writing them down in letters, edicts, proclamations, and so on.
In some places, literacy never seems to have moved beyond the circle of trained scribes. In Greece, all of the existing examples of Linear B are written by just a few dozen hands - when the palaces in which the scribes worked fell into ruin during the late bronze age collapse, the writing system became completely forgotten. But in Babylon, writing did eventually spread beyond the community of elite bureaucrats.
Literacy rates in ancient times is a vexed question - the vast majority of people never learned to read, but that's the case in every era up until the early modern period's expectation that all children should attend school before being put to work. But there is a huge difference between a literacy rate of much less than one percent - as for Linear B writers in bronze age Greece - and and literacy of a few percent - enough for there to be a tradition of vernacular writing, enough for the preservation of writing technology to be immune to changes or collapses of government.
Most of the archaeology that's been done in Mesopotamia, especially in the 19th and early 20th centuries, has focused on palaces and the homes of the elite. But some ordinary urban homes have been excavated, and Veldhuis cites a survey of the clay tablets found in those non-elite homes which found a fair number of tablets in the homes of ordinary city dwellers in ancient Babylonian cities, showing that by the second millennium, literacy had spread beyond graduates of the scribal schools to urban people of relatively modest means.
In Egypt, the story is more complicated.
The start of Egyptian writing is still murky because Egyptians lived on the flood plain of the Nile - the remains of most ancient towns are deeply buried under alluvial sediments and unfindable by archaeologists. Plus the damp soil by the river works against preservation of artefacts other than stone and pottery. Away from the river was desert where very few people lived - but people went there to worship and bury their dead. It's easy to find and excavate things in the desert, and perishable items like papyrus have survived well there, but with a tiny list of exceptions, you're not digging up places where people actually lived (exceptions include towns where quarry workers and tomb builders lived, and a few waypoints along roads built through the desert to connect cities on the Nile where the river was too curved to be the fastest route).
Aside from those exceptions, which have yielded graffiti and day to day writings on papyrus and ostraka (pot shards used for note taking), our knowledge of writing in Egypt is filtered through the items that were buried with dead people or inscribed on temples/monuments. I can't count the number of times I've seen someone say that ancient Egyptian culture was obsessed with death and the afterlife, and, no, that is a mistaken impression based on the overwhelming majority of the writing we've found being funerary.
The earliest preserved writing yet found, from the 4th millennium, appears as brief labels on grave goods - something like warehouse inventory tags. A bit later, murals in tombs start to have captions (although some tombs have a place for a caption but nothing written there - writing was ancillary and supplemental in tombs for a long time. The Great Pyramid, famously, contains no murals or inscriptions at all). The limited evidence found suggests a gradual evolution similar to cuneiform, from a writing system used mainly as an administrative tool in the 4th millennium to one that could record literature, letters, and biographies in the mid 3rd millennium. While pre-literate Sumer and Egypt were aware of each other and traded, there's zero evidence of technology transfer WRT nascent writing systems (scholars continue to argue over whether Egyptian writing was inspired by Sumerian writing, but as new finds continue to push the date of the earliest known hieroglyphs further back, it looks more and more like the amount of inspiration was extremely minimal to non-existent).
Very quickly after Egyptian writing moved beyond the warehouse and account books, it became a marker of prestige and power. Elites in Egypt often included the fact that they were a scribe in their biographies or official titles. Not that they spent a lot of time doing scribal work - that was relegated to clerks - but that they had the skill. The *fact* that they could read and write enhanced their power and prestige. So literacy became near-ubiquitous among rulers, priests, and the rest of the 1%.
Elites saying "this is our thing" helped give writing a very different path than in Mesopotamia. We have some glimpses into literacy amongst ordinary people - such as papyrus and ostraka found in Deir el-Medina, the village of excavators and artisans who built tombs in the Valley of Kings from 1500-1000. But the overwhelming majority of preserved writing was produced by or for the ruling class, and they liked keeping their toy to themselves. One tool of exclusion was the way the writing system did not evolve to keep pace with changes in the spoken language.
From 2500 to 500-ish, you had a very stable writing system - but in that time the spoken language transformed radically. Egyptian language evolved from synthetic (prefixes and suffixes indicate grammar) to analytic (helper words indicate grammar), so the whole structure of the language had changed. Plus, of course, evolution in pronunciations meant that many of the rebuses built into the sign system no longer worked as clues to how to interpret the signs. Imagine if people spoke English, but it was impossible to write anything except in classical Latin? Like that. Numerous examples of monumental inscriptions have been found where the scribe slipped up and produced a text that was more like how they spoke than how they were supposed to write.
Technical side note: Egyptian writing had two forms, hieroglyphs which were used for inscriptions on stone in tombs and on monuments, and hieratic writing, which was a simplified version of the hieroglyph signs optimized for handwriting on papyrus. These started out as merely different fonts of the same writing system in the same language.
Around 1350 BCE, people finally cracked under the strain of not having a writing system for the language they spoke, and in hieratic documents, especially literary works, you get Late Egyptian, with the same sign system but using analytic grammar, changes in the consonantal values used for writing words, and generally shifts which (scholars assume) moved the writing system closer to how people were talking. At the same time, in hieroglyphic inscriptions, Late Egyptian signifies a different development, with a huge explosion of new signs and much more complicated texts.
Another side note: Hieroglyphs had always been very pictorial, consisting of easily recognized little drawings of people, animals, plants, etc. To recap, each sign had a sound meaning (often but not always based on a rebus of the word for the thing depicted) or a determinative meaning (again not always related to the thing depicted), but it could also have a logographic meaning (yet again not always clearly connected to the depiction). In short, the system lent itself to cryptic writing - deliberately choosing obscure signs, seldom used logograms, complicated consonantal signs instead of common ones, and so forth, to produce texts that took extra effort to read. Taking it a step further, you could easily create complications to the system to write in code, using special rules that made your text unreadable to someone not in the know.
Late Egyptian hieroglyphic inscriptions doubled down on the cryptic potential of hieroglyphs, adding thousands of new signs in the process and generally going out of its way to be harder to read - and while no sources I found talk much about *why* this was done, it might have been a reactionary response to increased literacy among non-elites, a desire to keep the system "ours" and not let the hoi polloi have it.
And then there was another upheaval around 500 BCE, and hieratic writing ceased to be used except in religious contexts. Instead an even more simplified handwritten sign system called demotic emerged, which was used for everything else. Demotic altered the sign system substantially, again bringing it closer to how people spoke (which had evolved since the 1300s) as well as simplifying the individual signs to be even easier to write. Scholars today can connect demotic signs to their pictorial forebears, but looking at them in isolation, they were no longer recognizable little pictures of things - which helped since those pictures hadn't been a useful guide to meaning for a long time. I haven't found any explicit discussion of the number of signs needed to write demotic, but a presentation by Irena Ristevska on youtube includes a slide saying her team needed about 500 glyphs to create a digital font of the demotic script on the Rosetta Stone.
Returning to the question of why neither cuneiform nor hieroglyphs/hieratic took over the world the way alphabets did:
On the one hand, both scripts were at least somewhat burdened by complexities of adaptation that alphabets did not suffer from nearly as much. Syllabaries do not travel well - you have to reinvent the syllabary for each new language you port the writing system to. Isolated consonants and vowels vary a lot less between languages. As for Egyptian's system, the rebuses and the pictorial nature of the script became a hindrance, not a help, to someone trying to adapt the system to a new language. Egypt had the 24 single consonant signs from the beginning, but no one ever took those as the basis for a simplified, stripped down system the way the Hittites and Elamites did with cuneiform.
On the other hand, the biggest factor seems to have been the contingencies of history. A long-lived, incredibly stable scribal education system and culture in Mesopotamia created a pool of scribes who would happily translate Akkadian writing into languages understood by their employers, vastly reducing the need to adapt the writing system to the local language in lands on the periphery of Mesopotamia. Which largely cut off demand for adapted, localized writing systems in peripheral kingdoms - and that lack of adaptation made it hard for the technology to spread out beyond the sphere of Akkadian/Babylonian/Assyrian influence to lands outside the ken of Mesopotamia's scribal schools. The Hittites seem to be the farthest flung nation that used cuneiform.
Geography was against Egypt's writing system spreading far beyond the Nile - with deserts to the east, west, and northeast, the only close neighbours Egypt had were on the upper reaches of the Nile or its tributaries. Kush (modern Sudan) began using hieroglyphs around 1000 BCE, after they conquered Egypt for a time. If ancient Ethiopia started using hieroglyphs, I've found no reference to it. Other than those two, Egypt's closest contacts were in the Mesopotamian sphere and already had a perfectly good writing system, TYVM. But most of all, Egypt's writing system was "owned" by elites who had an interest in keeping it complicated and exclusive.
As for alphabets:
Sometime before the end of the 19th century BCE, a group of non-Egyptians who spoke an early Semitic language, and who lived in or next to Egypt, devised a consonantal alphabet (an abjad) based on 25 or so Egyptian hieratic signs. The earliest examples of the script in question have been found in two places, one as graffiti at Wadi el-Hol, on the desert road between two Egyptian cities, dated to the 19th century BCE; and the other in a turquoise mine and in the temple beside the mine, in Sinai, which is known to have been actively worked in the 19th and 16th centuries. Called Proto-Sinaitic after the place where it was first found, the script clearly does not use Egyptian sound values for the signs. However, if you assume the script is for a Semitic precursor to Canaanite/Phoenician/early Hebrew (originally all the same language), and use those sound values, it's possible to read a few words. Thousands of very short, often fragmentary examples of Proto-Sinaitic have been found throughout the Palestine-Lebanon region, with various 2nd millennium dates. Nobody knows anything about who invented Proto-Sinaitic, and the extremely short, fragmentary finds make it unclear how and by whom it was being used over the course of the 2nd millennium, but it was clearly being used in the region and a line can be drawn from it to the emergence of written Phoenician around 1000 BCE (keeping in mind that while Phoenicia the maritime powerhouse was a specific group of city states in what is now Jordan, Phoenician culture, language and writing were mostly indistinguishable from those of the people living in Canaan and Palestine at the time).
During the 2nd millennium, the city-states that later became Phoenicia were important to Egypt only as a source of cedar wood and a link on the trading roads to Mesopotamia. Egypt conquered them but allowed them autonomy in exchange for wood and free passage to the east. As maritime powers go, the city states were nothing compared to the Mycenaean civilization in Greece and Crete. Then, in 1200, a poorly understood series of calamities dubbed the Late Bronze Age Collapse destroyed the Mycenaean civilization and badly weakened Egypt - and the Phoenicians stepped into the resulting power vacuum. They built ships and created a powerful trans-Mediterranean trading network, founding or conquering numerous North African port cities, and becoming dominant sea traders in the region until the re-emergence of Greece in the 9th century (note that they were never politically unified, so there was never a Phoenician empire, just several cities each with its own trading network). And after 1000, everywhere they went, Phoenicians took their alphabet with them. Would they have spread so far if they hadn't been the only source of Tyrian purple dye, made from those snails Incunabula talks about? Probably. The dye certainly helped, but fundamentally they were seafaring traders, and if we must point to a single emblematic regional product that powered their rise, it would be the cedars that they used to build their ships.
But that alphabet of theirs didn't just travel west - it also went east, and neither snails nor cedar had anything to do with that spread. The Ancient Aramaic alphabet emerged in the Persian empire around 800 BCE. Because Persia was large and had many languages, there was also the Old Persian cuneiform writing system - starting around 500, Persian speakers junked the syllabary/determinative/logogram edifice of the past three thousand years of cuneiform writing and instead started to use a small set of cuneiform signs as an alphabet. Meanwhile, Proto-Sinaitic developed into Ancient South Arabian also around 800, which in turn inspired alphabets in Ethiopia and the Horn of Africa.
In southern India, the Brahmi alphabetic script appears suddenly around the 6th-4th century BCE, with no easily traceable path back to alphabets from the west (Persia never spread further than Pakistan and the Indus valley; India traded with Arabia but there's no recognizable connection between the either Persian or West African/South Arabian scripts and Brahmi). And yet, there's no identifiable predecessor to Brahmi in India either (between the extinction of the undeciphered Indus valley writing system in 1900 and the emergence of Brahmi ~1500 years later, no sign of writing existing in India has been found). The question of whether the Indian alphabets were independently invented or owe their genesis to scripts from the west remains a vexed question for scholars, made more contentious by Indian nationalists, including some scholars, who very strongly *want* India to have invented its own alphabet from scratch without outside influence or inspiration. Whatever Brahmi's origins, writing systems based on it spread throughout India to Tibet, Southeast Asia, and eventually to Korea, where Sejong the Great, inspired by the Tibetan script, invented the Hangul alphabet in 1443 CE as a supplement or replacement for Chinese characters, which had been used in Korea up until that time but were ill-suited to spoken Korean (Chinese is analytic; Korean is an agglutinative synthetic language).
Sejong is thought to have been inspired by Tibetan writing, but he invented a unique system designed to be simple and fast for the common people to learn. The populist slant of Hangul was frowned upon by Korean nobility, and in 1504 Hangul was banned after the publication of a tract criticizing King Yeonsangun, but the ban did not last and with minimal modification Hangul continues to be used in Korea today.
Which brings us, finally, to printing and movable type. (sources: wikipedia, the Chinese sections of Jonathan Bloom's "Paper Before Print," some Reddit threads, google searches based on memories of a book I read in college 33 years ago, and a paywalled article, "Early Korean Printing" by Pow-key Sohnn.
Woodblock printing originated in China sometime before 220 CE and for a long time was mainly used to print pretty patterns on silk. Buddhists seem to have been the first to realize that you could use woodblocks to print many copies of writing on paper. They were not interested, at first, in doing this to to make books to be read. Mahayanan Buddhism regards religious texts, as physical objects, to have talismanic power, and the creation of such texts as a meritorious act. Woodblock printing made it possible to produce hundreds or thousands of textual talismans, and thus generate far more merit than was possible with hand copying a text. Buddhists began to use printing to churn out such not-intended-for-reading texts in the 600's. Some were distributed, but many more were simply ritually buried en masse after being made - the point was the making of them. But printed Buddhist texts that were meant to be read started to appear very soon afterwards - the earliest known is thought, based on analysis of the character forms, to have been printed at the end of the 600's. Non-Buddhist books began to be printed in significant numbers in the 800's, and printing started to become quite common in the 900's.
Woodblock printing spread quickly to Korea and Japan - In Korea the earliest printed book found dates to the early 700's; in Japan around 770 Empress Shotoku commissioned the printing of a million copies of a Darani sutra for talismanic purposes, with each copy encased in a miniature pagoda. Printing of books meant to be read was happening by the 900's in Korea and the 1000's in Japan.
Histories of printing almost always are actually histories of printed books, and neglect how printing transformed realms other than books. For a government bureaucracy, printing is obviously extraordinarily useful, allowing mass production of flyers, posters, forms, and so on. But I couldn't find anything online about when exactly the imperial government embraced printing, except with reference to imperial commissioned editions of books.
In the late 600's emperor Wu Zetian expanded and made permanent the nascent imperial examinations system. Aspiring examinees needed copies of the canonical Thirteen Classics of Confucian tradition to study, but seem to have had to make do with handwritten copies until the 930's, when the government commissioned a printed edition of those classics. Closer to the end of the 900's, the first few Song emperors supercharged the uptake of printing with a vast expansion of the examinations system. A text from 1005 describes a visit by the emperor to the imperial Publications Office, which had 100,000 woodblocks in its warehouse at the time, in contrast to the mere 4,000 blocks that had been in stock at the start of the dynasty's regime in 960. The publications officer went on to enthuse about how, thanks to printing, it was so much easier to obtain copies of the classics under the rule of the current emperor and his predecessor than it had been under the previous dynasty.
Handwritten books continued to be more prestigious than printed ones. Book collectors prized and desired handwritten books above printed ones, but printed books were *ten times* cheaper - which means they became far more abundant. In all of Chinese history before 960, a total of 200 private libraries of books are known, and only 140 of those are from the era before printing began to take off. In the next four centuries (the Song dynasty), we know of 700 private collections. The sizes of those collections increased dramatically after printing as well. Chinese retained the term "juan" (scroll) as a unit of book measurement even after the rise of codexes, which conveniently allows direct comparison between scroll based (pre 900's) and codex based (post 1000) libraries. The pre-printing library of 5th century scholar Shen Yue had 20,000 juan. In the 1100's, the scholar Ye Mengde had 100,000 juan of books totalling 6,000 titles.
The huge increase in the availability of books boosted literacy. In 1488, a ship of Koreans, including a government official Choe Bu, was blown far off course by a multi-day storm and found themselves on the coast of Southern China near Ningbo. As a government official, Choe Bu could read and write Chinese characters, and he found that despite not knowing a word of spoken Chinese, during the shipwrecked Koreans' long trip back home, he could communicate in writing with almost everyone he met - "even village children, ferrymen and sailors."
Woodblock printing technology started with writing the text normally on translucent paper, which was pasted upside down onto a wooden block. Then a wood carver would chisel away a layer of wood everywhere that the paper had no ink, leaving only a mirror image of the desired characters. If a mistake escaped the initial proofreading stages and got carved into the block, it could be chiselled out and replaced with an inlay containing the correct text. The finished block would be inked and a sheet of paper pressed firmly onto it. Since the block was not being smashed down onto the paper as is done with European-style printing, the paper could be relatively thin and delicate. The most time consuming part was obviously carving the wooden blocks; for big projects, storing and wrangling all the blocks (in case more copies were needed later) could get challenging - in Korea, a 971 edition of the Bhuddist Tripitaka scriptures used 130,000 blocks.
In 1040, a Chinese printer named Bi Sheng decided he was tired of dealing with so many blocks and invented movable type. He experimented with making the individual pieces of type from wood, but quickly discarded that in favour of ceramic. The scholar Shen Kuo wrote a detailed account of his invention and process, talking about how most words had "several" pieces of type for them while more common words had "twenty or more"; how all of these many thousands of pieces of type were organized by rhyme group with paper labels; and how Bi Sheng usually had one page being printed while he composited the next; and how "if one were to print only two or three copies, this method would be neither simple nor easy. But for printing hundreds or thousands of copies, it was marvellously quick."
Ceramic pieces of type had to be individually sculpted, and sometimes slumped or warped in the kiln. Making type from bronze, cast from moulds using wooden masters, produced much more uniform results. Paper money started being printed with bronze type in the 1100's. In 1190, Zhou Bida read Shen Kuo's account and made ceramic type for printing his book. In 1234, the first books made with metal type were printed in Korea. Around 1300, in China, Wang Zhen published a book on how to print books using wooden type. He also wrote about his system for sorting the type: he used a pair of rotating round (multi level?) tray tables on which compartments of type were sorted by tone and rhyme; the compositor could spin the tables to bring the next needed word into easy reach.
Wood, metal and sometimes ceramic type continued to be used to print some books in China, but woodblock printing remained the default for 90% of printing until the introduction of European mechanised printing methods. There were a few obvious reasons why movable type did not displace woodblocks in China. First, upfront costs: woodblock printing required wood and a wood carver, who did not need to be literate. Movable type required the creation of tens of thousands of pieces of type, minimum, before you could print a single sheet. Second, training: any carpenter could carve woodblocks. Movable type required setting up a complex organizational scheme for storing the type, and then a highly literate compositor had to learn that scheme by heart. Third, scale: setting movable type was more economical only for large jobs. For the small jobs which are the bread and butter of any printing business, it made more sense to just make woodblocks.
Things were a bit different in Korea, which did not have many hardwood trees suitable for making woodblocks: metal type became much more common in Korea for this reason, but printing remained a government monopoly, with independent printers and the selling of books strongly discouraged. So there wasn't the explosion of books and literacy that occurred in China. That government monopoly also prevented movable type technology from spreading to Japan until the late 1500's, by which time Jesuits from Europe had already arrived with printing presses of their own (Japan's relations with China were at arm's length due, among other things, to a long term problem with Japanese pirates preying on Chinese ships).
Incunabula claims that printers in East Asia experimented with movable type but gave up on it. Which is a common claim in Eurocentric histories, but it's not true. Movable type became a known technology, and a steady trickle of books were produced with it in China - but only when it made financial sense to do so. For most jobs, making woodblocks was cheaper. It was only for large jobs, mainly government commissions, that movable type made sense. Behind that steady trickle of movable type books was a small group of printers who made the heavy investment in type, memorized a storage system, and sold their services to the government. Movable type would have had a larger share of the printing market if China had a tradition of large print runs, but for whatever reason, small runs of a hundred or so copies were much more typical.
Printing and the Korean alphabet: In June of this year, a trove of bronze type was dug up in Seoul that included some of the earliest Hangul type yet found, from around 1460 - but the cache contained more Chinese type than Hangul type. The Hangul alphabet and movable type didn't get a chance to show their synergistic power in Korea for a long time, mainly because the 1% in Korea at the time were Sinophiles, so their desire was for printed editions of Chinese literature. For writing in Korean, the elites preferred to use Chinese characters supplemented by Hangul for the grammatical affixes and particles that did not exist in Chinese. The lower classes and women, who were the primary users of pure Hangul, were not able, or not allowed, to set up printing operations.
Printing also moved west. India started making cotton fabric prints in the 900's. Never having adopted paper, they did not start printing books there. In the Islamic world, block printing was mainly used for making charms and talismans; the earliest known, from Egypt, also dates to the 900's. Arabic script was entirely cursive, and movable type is really only suited to scripts with separation between the letters (the first Arabic type fonts that produced acceptable looking results were not made until the modern era). The same feeling that handwritten books were superior existed in the Islamic world as in China, only much stronger. Completely aside from the issues with making attractive Arabic type, calligraphy in the Islamic world was a highly prized art form, and there was zero interest there in producing cheaper yet less beautiful books by using woodblock printing.
Because of this, printing arrived in Christian Europe as a technology for printing patterns on fabrics, completely divorced from any idea that it was useful for making books. Paper arrived a little bit after printing, and by the 1400's, printing started being used with paper to produce art prints, playing cards, and such. The earliest wood block printed European books were heavily illustrated, and were being made by 1440 if not earlier. From that as a starting point, either Gutenberg had the same epiphany as Bi Sheng five hundred years before, or someone who had been to China told him about movable type. Nobody knows which it is - Gutenberg wrote nothing about his invention, himself, or his printing shop, and most of what is known about the press he made is inferred from copies of his Bible and from other early printers who copied his setup, and did write about their tools and procedures. He does seem to have been the first person to reverse the relation of the type/block and the paper, using a screw to press the inked type down onto the paper rather than pressing the paper onto the inked surface.
In Europe, the movable type printing press brought about a massive series of cultural changes which ended the medieval period, splintered a monolithic Christianity into competing sects, enabled a rise in scholarship and powered a series of revolutions in science. There's a tendency among European historians to say that it was the synergistic combination of an alphabet and movable type that made all that possible, but looking at China, I don't think that's right. And there's another trend to say that China and Korea had strong unified governments that could tell printers and scholars what to do and what to study, whereas 15th century Europe had tiny fragmented polities ruled by incredibly weak kings, and that enabled more intellectual freedom, less censorship, and thus more technological and cultural experimentation and change. But I don't think that's it either.
Europe in the 1400's was incredibly poor and powerless, but when printing was still just beginning there, Europeans were already working to enrich themselves, and were already travelling down the path that led them to rapacious conquest, enslavement, colonization, and general global bullying. They had already started to have a period of intellectual ferment and technological change based on rediscovery of lost classics and new innovations, and that ferment was being fuelled in part by greed and the desire to devise better tools for warfare. Printing just pushed things along faster.
When post-Gutenberg Europeans first arrived in China and looked around, they (mostly very bigoted Jesuit missionaries) saw a culture where books were not just ubiquitous, but actually as cheap or cheaper than they were in Europe. They spoke of that as a bad thing (because the books were Confucian or Buddhist, and they were Christian bigots). But my takeaway is that China didn't need movable type to have an explosion of literacy, of books, and of scholarship. It had already had its own printing revolution, about six hundred years before Europe's. But that revolution expressed itself in a very different form, with very different long term results. China was already rich, already powerful. Instead of a quest for wealth and power that in turn fuelled fast scientific and technological developments, printing in China powered a quest to keep corruption in check and the exploitation of the poor within acceptable limits, by instituting government by those who had deeply studied Confucian principles.
Looking at the world we've gotten thanks to technology, science, and rampant exploitation fuelled by European greed, I can't say China was foolish to take the path it did.
Bit that didn't fit above:
The Dunhuang Manuscripts are a cache of scrolls and codexes found in 1900 by a Buddhist monk who was looking to rehabilitate a disused monastery in Western China. Totalling about 500 cubic feet of paper, the manuscripts had been sealed inside a cave in the early 1000's. They're an incredibly rich source of historical information, as well as a treasure trove of knowledge about Chinese paper and book making from the 4th to 10th centuries. Sadly, the dysfunctional nature of China's government of the early 20th century meant that when the discovery was reported to authorities, they hemmed and hawed about the expense of warehousing and moving so many books back east for conservation and study. After a few years of getting nowhere, the discoverer of the cave, needing money for his project, sold many of the manuscripts to European looters.
The usual history of Chinese bookmaking posits a neat, orderly succession of formats from scrolls through various refinements of the codex - pasting the pages together at a spine, wrapping the spine, and finally sewing the pages instead of gluing them. Looking at examples from Dunhuang that ended up in the British Library, Li Zhizhong's "Problems in the history of Chinese Bindings" shows how instead of an orderly progression of innovation, all of those formats were experimented with in the 900's before the "butterfly binding" won out in the 1000's.
The codex format spread from Rome to Persia and Arabia but foundered in the steppes and deserts between Persia and China. The earliest paper books in China mimicked the scroll format set in the era of writing on bolts of silk: manuscript pages written on one side were pasted together, then rolled up. All of the pre-900's printed books which survive are scrolls. A scroll got unwieldy with longer manuscripts, and in the 900's book makers seem to have tried various alternative formats which can be found in the Dunhuang cache, and all of which are spoken about in Chinese histories, but as representing a succession of formats. "Whirlwind" or "dragon scale" binding was a compromise adaptation of the scroll format: it pasted individual manuscript pages (written on both sides) to a relatively short blank scroll, which could then be rolled up and stored like a scroll. Unrolled, you had a long strip of overlapping pages, looking rather like scales, which could be turned through. The next step was to paste the pages much closer together onto a much smaller backing paper, producing something halfway in between a whirlwind scroll and a codex - a format found in Dunhuang but not attested to in historical records. The final form that caught on for the next few centuries was "butterfly" binding - which dispensed with the scroll entirely and directly pasted a stack of folded pages together along the fold, producing a codex that fell open like a butterfly's wings. Since Chinese woodblock printing usually used thin paper printed on only one side, you had a pair of blank pages between every pair of printed pages. The butterfly format had staying power and remained very popular from 1000 through the 1300's, when printers started using "wrapped back" binding. This reversed the glue point to the edges of the folded sheet, hiding the blank side of the paper within the fold. Gluing a backing onto the spine helped strengthen the book and guarded against the pages falling apart from glue failure. Finally, "wrapped back" books started being sewn instead of glued in the 1400's... but again, sewn manuscripts were found in the Dunhuang cache. All of those book formats were printed on very thin paper, so Chinese books tended to be quite floppy and delicate. Sometimes they were strengthened with stiff covers; other times each book was stored in its own small box.