When Was Writing Invented?

The oldest writing dates to around 3400 to 3200 BCE in Mesopotamia, and it was invented from scratch at least four separate times

The earliest known writing was invented in southern Mesopotamia around 3400 to 3200 BCE, in the city of Uruk, in the form of Sumerian proto-cuneiform pressed into clay tablets. Egyptian hieroglyphs appear at very nearly the same moment, with inscribed labels from Abydos dated to roughly 3400 to 3200 BCE, which is why no serious scholar declares a clean winner between the two. Writing was then invented independently at least twice more: in China, where the oracle bone script is attested by about 1250 BCE, and in Mesoamerica, where identifiable scripts appear in the last centuries BCE.

That is the short answer, and almost every word of it needs a qualifier. The dates carry a 'c.' for circa because they rest on stratigraphy, radiocarbon ranges, and the accidents of what happened to survive in the ground. The earliest tablets are also not obviously writing in the way a letter or a poem is writing. They are accounting records: counts of grain, beer, livestock, and labour, made by administrators who needed to know what had come into a temple storehouse and what had gone out. Whether those first tablets qualify as writing at all depends on a definition that specialists argue about, and that argument is the real content of the question.

So the honest formulation is this. If writing means a system of visible marks capable of recording spoken language, the Mesopotamian system crosses that line gradually across several centuries, with clearly syllabic, language-recording tablets appearing after about 2900 to 2800 BCE. If writing means the first durable, standardised, rule-governed symbol system used to store information, the answer is Uruk, around 3400 to 3200 BCE. Both answers are defensible, and the difference between them is the difference between proto-writing and writing.

Where the oldest writing comes from, and what it says

The oldest securely dated documents come from Uruk in southern Iraq, from the archaeological level known as Uruk IV, with the first unequivocal written documents placed at around 3300 BCE and precursor numerical tablets earlier still. Uruk at this date was one of the largest settlements on earth, and the tablets belong to the machinery that ran it. They record quantities of barley, textiles, beer rations, and animals, using a combination of numerical signs and pictographic signs for the commodity being counted.

This matters for the popular version of the story. Writing was not invented by a poet, a priest with a revelation, or a king who wanted his name remembered. It was invented by bureaucrats, and its first job was inventory control. Literature, law, and history all arrive later, using a technology built for stock-taking. The earliest recognisable literary compositions in Sumerian are separated from the first tablets by something like six to eight centuries.

There is also a deep prehistory to the medium. For thousands of years before the first tablets, people in the Near East used small clay tokens of standardised shapes to represent commodities, sometimes sealed inside clay envelopes with impressions of the tokens marked on the outside. The transition from impressing a token into clay to drawing the sign for that token onto a flat tablet is short and mechanical, which is one reason the invention looks less like a flash of genius than like the last step in a long administrative habit.

Cuneiform, the wedge-shaped script that grows out of proto-cuneiform, then had one of the longest careers of any writing system. It was used for Sumerian, Akkadian, Hittite, Elamite, and other languages, and the last known cuneiform inscription, an astronomical text, dates to 79 or 80 CE. That is well over two and a half thousand years of continuous use.

Proto-writing versus writing, and why the distinction decides the date

Proto-writing is a system of ideographic or mnemonic marks that conveys information without recording a specific language. A tally, a set of counters, a picture sequence that reminds a reader of a story he already knows: these carry meaning, but a reader who does not already know the content cannot reconstruct the exact words. Full writing is different. It encodes enough of the sound and grammar of a particular language that a competent reader can recover an utterance he has never heard before, including a name he has never seen.

By that test, the earliest Uruk tablets sit on the boundary. They are unambiguously a system, standardised across sites and stable over generations, but they do not spell out Sumerian sentences. The decisive move is the rebus principle, where a sign for a thing gets reused for the sound of that thing's name, which lets a scribe write words that cannot be drawn, including personal names and grammatical endings. Signs begin taking on phonetic values in the centuries after 3000 BCE, and by roughly 2800 BCE tablets show syllabic elements and the grammatical structure of the agglutinative Sumerian language.

This is why you will see the invention of writing dated to 3400 BCE, 3300 BCE, 3200 BCE, 3100 BCE, or even 2900 BCE depending on which book you open, without any of those authors being wrong. They are answering slightly different questions. The polite convention is to say that writing emerged in Mesopotamia in the late fourth millennium BCE, which is broad enough to be true under every definition.

Egypt, China, and Mesoamerica: four inventions, not one

Egyptian hieroglyphs are the second-oldest system, and the gap between them and Mesopotamia is small enough that the direction of influence, if there was any, is genuinely unresolved. Early glyphs from Abydos are dated to between roughly 3400 and 3200 BCE, and the developed hieroglyphic system is conventionally placed around 3250 BCE. Some scholars argue for stimulus diffusion, meaning Egyptians heard that Mesopotamians were doing something like this and built their own version rather than borrowing signs. Others argue for wholly independent invention. The two systems share no signs and work on different principles, which is the strongest argument for independence.

China is a clearer case. The earliest attested Chinese characters are the oracle bone inscriptions of the late Shang period, around 1250 BCE, carved onto turtle plastrons and ox scapulae used in divination. The script is already mature when it appears, which implies an earlier developmental phase that has not survived, probably because it was written on perishable materials. There is no plausible route by which Mesopotamian or Egyptian writing reached the Yellow River valley in a usable form, and the internal logic of the Chinese system is distinct. It is treated as an independent invention.

Mesoamerica is the fourth case and the least contestable, since the Americas were culturally isolated from Eurasia. Identifiable script appears in the last centuries BCE, with early Maya writing attested from around the third century BCE and clearly readable dated texts by the early centuries CE. Nobody argues for contact with the Old World here, so Mesoamerican writing is the cleanest proof that writing is something human societies can invent whenever record-keeping pressure gets high enough.

Every other script in the world descends, directly or by inspiration, from one of these four. The alphabet itself is a late arrival and a derivative one, growing out of a Semitic adaptation of Egyptian signs in the second millennium BCE and reaching Greece, with vowels added, in the early first millennium BCE.

Why the dates keep moving, and what would change them

The dates rest on excavation, and excavation is unfinished. Most of southern Iraq's fourth-millennium levels lie under later cities, under water tables, and under decades of war and looting. A single well-stratified deposit at an unexcavated site could push the earliest tablets back by a century or two. This is not a hypothetical worry: the accepted date for the earliest Egyptian glyphs moved earlier when the Abydos tomb material was published, and similar revisions could happen again.

Dating methods also carry their own uncertainty. Radiocarbon gives a probability range, not a year, and a tablet is dated by the layer it was found in rather than by itself. Where a tablet was reused, redeposited, or dug up in the nineteenth century with poor records, the date attached to it can be soft. This is the substantive reason to distrust any source that gives you a single precise year for the invention of writing.

There is a final wrinkle worth knowing. Several undeciphered or partly deciphered systems sit near the boundary, including the Indus script of the Bronze Age Indus Valley, which appears on thousands of seals from roughly 2600 BCE onwards. Specialists are divided on whether it encodes a language at all or functions as a non-linguistic sign system. Until it is deciphered, or shown to be undecipherable, it stays out of the count of confirmed independent inventions.

Frequently asked questions

When exactly was writing invented?

The earliest known writing is Sumerian proto-cuneiform from Uruk in southern Mesopotamia, dated to roughly 3400 to 3200 BCE, with the first unequivocal written documents placed around 3300 BCE. Egyptian hieroglyphs appear at close to the same time. No single precise year can be given, because the dates depend on archaeological layers and radiocarbon ranges rather than on any record of the event.

Was writing invented in one place and copied everywhere else?

No. Writing was invented independently at least four times: in Mesopotamia around 3400 to 3200 BCE, in Egypt at nearly the same date, in China by about 1250 BCE, and in Mesoamerica in the last centuries BCE. Whether Egypt was influenced by news of the Mesopotamian invention is debated, but China and Mesoamerica are treated as clearly independent.

What is the difference between proto-writing and writing?

Proto-writing uses ideographic or mnemonic marks that carry information without recording a specific language, so a reader cannot reconstruct exact wording. Full writing encodes enough sound and grammar that a reader can recover a sentence he has never heard before. Mesopotamian signs began taking phonetic values after about 3000 BCE, and tablets showing the grammar of spoken Sumerian appear by roughly 2800 BCE.

What was the first thing ever written?

Accounting. The earliest tablets from Uruk are administrative records of commodities such as barley, beer, textiles, and livestock, produced by the officials who ran the city's storehouses. Literature, law, and historical narrative appear centuries later, using a technology originally built for inventory control.

Why do dates for the invention of writing always have a 'c.' in front of them?

Because they are estimates. A tablet is dated by the archaeological layer it was found in, and radiocarbon dating produces a probability range rather than a year. Excavation is also incomplete, especially in southern Iraq, so a new find could move the earliest date. The 'c.' is a signal that the figure is a considered approximation, not a record of a known event.

Further reading

Prefer listening? History is Audible’s strongest catalog — an annual Premium Plus membership is the best value, or compare the monthly plans.

Explore related events on the timeline

• The Invention of Cuneiform on the interactive timeline
• The Epic of Gilgamesh on the interactive timeline
• The Code of Hammurabi on the interactive timeline

Related guides

• Best Books on the History of Writing
• Best Books on Mesopotamia and Sumer
• Best Books on Ancient Egypt

Sources consulted

• Wikipedia, 'History of writing'
• Wikipedia, 'Cuneiform'
• Open Library record: Andrew Robinson, The Story of Writing
• Open Library record: Barry B. Powell, Writing

Researched with AI and verified against mainstream scholarship and the cited sources on this page. See our methodology and corrections policy.

As an Amazon Associate, AskHistoryAI earns from qualifying purchases. Books are chosen editorially.