Writing was invented as a redundant verification copy of an accounting record

the problem

A sealed accounting record couldn't be checked without breaking the seal that made it trustworthy

background

what everyone would do

The obvious fix for 'you have to break the seal to check it' is procedural, not representational: keep a separate written ledger of what each bulla contains, or simply trust the parties' memory of what was sealed inside and reserve breaking the seal only for outright disputes. Both fail at scale — a separate ledger just relocates the tamper problem to a new document, and trusting memory in a dispute defeats the entire reason tokens were sealed in the first place.

what they saw

Whoever started pressing the tokens into the bulla's wet surface before sealing them saw that the verification mark didn't have to be a separate record kept elsewhere — it could live on the very surface of the tamper-evident container, readable without opening it. The insight that mattered even more was the one that came after: once the impressions carried every bit of information the tokens inside did, the tokens themselves had become pure redundancy.

the move

Sometime after roughly 3300 BCE, Mesopotamian scribes began pressing the tokens into the wet clay of the bulla's exterior surface before sealing them inside, leaving an external impression that showed exactly what was sealed within without requiring anyone to break it open to check. Eventually, according to archaeologist Denise Schmandt-Besserat's influential reconstruction, someone recognized that the external impression alone already carried all the necessary information — the tokens sealed inside had become entirely redundant — and dropped them, drawing the marks directly onto a flat clay tablet instead of enclosing physical tokens at all.

why it works

Pressing each token into the clay before enclosing it creates an external, tamper-evident record of exactly what was sealed inside, so a claim can be checked by reading the surface instead of destroying the seal — solving the original problem without giving up tamper-evidence. But because the impression is a complete, lossless copy of the information the tokens carried, the physical tokens add nothing once the impression exists; dropping them and drawing the marks directly onto a flat tablet keeps all the verifiable information while shedding the cost of manufacturing, sorting and enclosing physical counters. The verification copy, built only to let people check the original without disturbing it, turned out to be sufficient on its own — the pattern that produces writing here is the same one that shows up whenever an audit trail becomes more useful than the thing it was built to audit.

the payoff

Those impressed and later drawn marks evolved into proto-cuneiform and then cuneiform writing, giving the ancient Near East a system for recording information in an abstract, two-dimensional symbolic form rather than as physical, three-dimensional tokens standing in for real objects.

where it breaks

The move only works when the verification layer is genuinely a complete, lossless representation of the original — a summary, a hash, or a partial impression that drops information can't safely replace what it was meant to verify. It also depends on nobody's downstream trust resting on properties only the physical original has (a wax seal's material authenticity, a coin's metal content, a signature's ink) — where the substrate itself carries part of the proof, discarding it loses something the copy cannot replicate. And the transition only happens if someone notices the redundancy at all: an audit trail can sit unexamined for millennia, as archaeological evidence suggests these impressed bullae did for centuries, before anyone recognizes the copy no longer needs the original.

what came after

Schmandt-Besserat's tokens-to-writing thesis, while debated and refined by subsequent scholars, remains a foundational and widely cited theory in the archaeology of writing, treating the origin of written language not as a deliberate invention of a new communication technology but as the byproduct of an accounting verification step that turned out to be more useful than the object it was originally built to verify.

references

1. [1]From Accounting to WritingDenise Schmandt-Besserat, University of Texas at Austin, 2014sites.utexas.edu
2. [2]The Earliest Precursor of WritingScientific American, 1978scientificamerican.com
3. [3]Tokens, Writing and (Ac)counting: A Conversation with Denise Schmandt-Besserat and Bill MaurerExchanges: The Warwick Research Journal, 2017exchanges.warwick.ac.uk
4. [4]Proto-cuneiformWikipedia, 2025en.wikipedia.org

keep it

same kind of clever

• 1935 · BoeingAfter its own best test pilots crashed a plane by forgetting a single step, Boeing didn't train pilots harder — it wrote the steps down.
• 2020 · Burger King (Moldy Whopper campaign)Burger King proved its Whopper had no artificial preservatives by filming one decay into mold over 34 days and running the footage as the ad
• 2006 · Cedars-Sinai Medical CenterCedars-Sinai got doctors to wash their hands by photographing the bacteria growing on their own palms