On writing as the foundational technology of civilisation, from clay tablets to neural networks.

I. The Mark

Before writing, every thought was a flame held in cupped hands — beautiful, momentary, extinguished by the first wind.

A hunter in the Palaeolithic could look upon the stars and intuit something vast. A mother could whisper wisdom to her daughter beside a dying fire. An elder could speak of floods and migrations, of seasons learned through suffering, of which berries poisoned and which sustained. But when the elder died, much of what he carried died with him. Knowledge was hostage to memory, and memory was hostage to mortality. Every generation began, in some essential sense, almost from the beginning.

Then, roughly five thousand years ago in the alluvial plains of Mesopotamia, someone pressed a reed stylus into wet clay and made a mark that stayed.

It is difficult to overstate the magnitude of this act. Not fire — that was mastery over energy. Not the wheel — that was mastery over distance. Writing was mastery over time. The mark on the tablet did not decay with the body that made it. It did not distort with each retelling. It did not require the physical presence of its author to convey its meaning. For the first time in the long, brutal chronicle of our species, thought could outlive the thinker.

The dead could now speak to the living. And the living could answer.

This was the fracture — the moment the chain of forgetting was broken. Before writing, human knowledge was a river that evaporated as fast as it flowed. After writing, it became a reservoir. Shallow at first. A few accountants’ tallies, a few inventories of grain and livestock. But a reservoir nonetheless, and one that would only deepen.

Everything that follows — every cathedral, every constitution, every equation, every line of code — is a consequence of that first mark in the clay.

II. The Sacred Word

The earliest uses of writing were prosaic. Tallies. Transactions. The bureaucratic minutiae of emerging city-states. But it did not take long for the scribes to recognise that a tool capable of recording bushels of barley was equally capable of recording the words of the gods.

And so writing found its first great vocation: the preservation of the sacred.

Oral traditions, once fluid and mutable — reshaped by each teller, adapted to each audience, softened or sharpened by the vagaries of memory — were fixed into text. The Epic of Gilgamesh. The Vedas. The Torah. What had been living narrative, breathing and shifting with each generation, became canon. Permanent. Authoritative. Unchallengeable.

The consequences were immense. With written scripture came a new kind of power: the power of interpretation. If the word of the divine was inscribed on tablets, then whoever could read those tablets — whoever controlled access to the written word — controlled the divine itself. The scribe became priest. The library became temple. Literacy became a form of ordination, restricted to the select few who could be trusted not to mishandle the sacred.

Writing, which had liberated thought from the prison of individual memory, now constructed a new prison: the prison of orthodoxy. The spoken word invites dialogue; it is inherently democratic, for anyone with a voice may contribute. The written word, by contrast, invites reverence. It sits upon the page with an authority that speech can never muster, and that authority, once established, resists all revision.

The irony is exquisite. The very technology that would eventually enable philosophy, science, and the dismantling of superstition was first employed to render superstition permanent. The sacred texts that would dominate human consciousness for millennia were writing’s first masterworks — and its first chains.

Yet even chains, if forged well enough, become foundations. The sacred word, for all its tyranny, accomplished something indispensable: it taught humanity to take text seriously. To treat written symbols as bearers of meaning worthy of study, of commentary, of generation upon generation of devoted attention. This reverence for the written word — initially directed toward the divine — would in time be redirected toward objects more deserving of it.

III. The Rational Line

Somewhere in the sixth century before the common era, along the Ionian coast, a subtle but decisive shift occurred. Men began to write not the words of gods, but the words of reason.

Thales of Miletus proposed that the world was composed not of divine whim but of natural substance. Pythagoras discovered that the intervals of music could be expressed as numerical ratios — that harmony itself was mathematical. Heraclitus wrote his fragments on the nature of change. These were not revelations received from above; they were arguments constructed from below, offered not as commandments but as propositions, subject to examination, to refutation, to improvement.

This could not have happened without writing.

Philosophy is, at its root, a written enterprise. The spoken argument vanishes into the air; the written argument persists on the page, available for scrutiny by minds not yet born. When Plato set down his dialogues, he created not merely a record of Socratic thought but a mechanism for its perpetuation and critique across millennia. When Euclid composed his Elements, he demonstrated that complex truths could be derived from simple axioms through rigorous logical steps — each step written, each step verifiable, each step a foundation for the next.

Mathematics, in particular, is writing’s most extraordinary child. It is a language of pure abstraction — symbols manipulating concepts that no voice could articulate, no gesture could convey. The equation $e^{i\pi} + 1 = 0$ unites five fundamental constants in a relationship of breathtaking elegance, yet it could not exist outside the written form. It cannot be spoken without losing its precision. It cannot be remembered without the symbolic apparatus of notation. It lives only on the page — or, more precisely, it lives through the page, in the minds of those who have learned to read its particular dialect of the written word.

What writing made possible, then, was not merely the preservation of knowledge but its accumulation. Each generation could begin not from the beginning but from the summit of all previous thought. Aristotle could build upon Plato, who built upon Socrates. Archimedes could extend Euclid. Newton could extend them all. The written record transformed knowledge from a circle — endlessly rediscovered and endlessly lost — into an arrow, pointed forward, gathering momentum with each passing century.

This is the true revolution. Not fire. Not agriculture. Not bronze or iron or steam. The written word, and the cumulative edifice of thought it made possible, is the singular engine of civilisation. Everything else is a downstream effect.

IV. The Accumulation

For three millennia, writing remained the province of the few. Manuscripts were copied by hand, painstakingly, one at a time, in monasteries and scriptoria where a single volume might represent years of labour. Knowledge accumulated, yes — but slowly, and within narrow channels. The reservoir deepened, but its banks were high and closely guarded.

Then, in the fifteenth century, a goldsmith in Mainz constructed a device that would breach those banks forever.

Gutenberg’s press did not invent writing. It did something arguably more consequential: it made writing scalable. A text that had previously required months to reproduce could now be produced in hundreds, then thousands. The written word, once the exclusive currency of priests and scholars, flooded into the hands of merchants, artisans, and eventually the common literate citizen. Luther’s theses. Copernicus’s heliocentrism. Vesalius’s anatomies. Ideas that would have taken generations to propagate through manuscript culture now spread across continents within years.

The industrial revolutions that followed — each one a seismic reorganisation of human capability — were, at their foundation, revolutions of written knowledge applied. The steam engine did not emerge from intuition alone; it emerged from Boyle’s documented gas laws, from Newcomen’s published designs, from the accumulated correspondence of natural philosophers who shared their findings through the medium of print. The electrical revolution drew upon Maxwell’s equations — written equations, published and disseminated, available to any engineer literate enough to read them. The chemical revolution drew upon Mendeleev’s periodic table, Lavoisier’s nomenclature, Dalton’s atomic theory — all of them fruits of the written tradition.

Newton’s famous declaration that he had seen further by standing on the shoulders of giants is, in truth, a declaration about writing. The giants in question were not present in the room. They were present on the page. Their shoulders were made of paper and ink. Without writing, there would be no shoulders to stand upon — only the flat ground of each generation’s unaided perception, and the long, pointless labour of rediscovering what had already been discovered and forgotten a hundred times before.

Each industrial revolution accelerated the cycle. More written knowledge produced more technology, which produced more capacity for writing, which produced more knowledge. The printing press begat the scientific journal. The scientific journal begat the research institution. The research institution begat the technical manual, the patent filing, the engineering specification. By the twentieth century, the accumulated written output of humanity had become so vast that no single mind could comprehend even a fraction of it.

The reservoir had become an ocean.

V. The Executable Word

And then the written word learned to act.

In 1936, a young mathematician at Cambridge named Alan Turing published a paper — a written document — describing an abstract machine capable of performing any computable operation. The machine operated on symbols written on a tape. It read them, wrote new ones, and moved according to rules that were themselves encoded in symbols. The entire apparatus was, in essence, a theory of text that does things.

This was writing’s most radical transformation. For five thousand years, the written word had been descriptive — it recorded, it argued, it instructed, it inspired, but it did not, in itself, act upon the material world. A blueprint described a bridge but did not build one. An equation modelled a trajectory but did not launch a projectile. There remained always a gap between the written instruction and its execution, a gap bridged by human hands, human judgement, human labour.

Code closed that gap.

A computer program is a piece of writing. It is composed of symbols arranged according to a grammar, structured by syntax, governed by semantics. It is authored, edited, reviewed, revised — all the verbs we associate with the literary enterprise. And yet it is unlike any writing that preceded it, because it is performative. It does not merely describe a process; it is the process. The text itself, when interpreted by the appropriate machine, becomes action.

Consider the implications. When a programmer writes a function that sorts a list, the written text does not instruct a human to sort the list — it sorts the list. When a systems architect writes the code that routes packets across a global network, the written word is not a metaphor for communication — it is communication. The word has become deed. The map has become the territory.

Software, if one steps back far enough to see it clearly, is the most consequential body of literature in human history. It runs the financial markets, guides the aircraft, regulates the power grids, mediates the conversations, stores the memories, and — increasingly — makes the decisions that shape the daily experience of eight billion human lives. All of it written. All of it text.

The stylus in the clay. The quill on the parchment. The fingers on the keyboard. The gesture is the same. Only the consequences have changed — by orders of magnitude that the Sumerian scribe could not have imagined, and that we ourselves can barely comprehend.

VI. The Dreaming Text

And now the written word has done something that no one — not Turing, not the Sumerian scribes, not the most feverish imaginations of speculative fiction — quite anticipated.

It has begun to write itself.

The large language models that have emerged in the first quarter of the twenty-first century are, stripped of all mystification, statistical architectures trained on the corpus of human writing. Every book, every article, every forum post, every scrap of digitised text that civilisation has produced — these are the training data, the raw material from which the model distils its patterns. The machine reads the totality of the written word and produces, from that reading, new text that is coherent, contextually appropriate, and — at its best — genuinely illuminating.

This is not mere mimicry. The neural network does not copy; it abstracts. It learns not specific sentences but the deeper structures that generate sentences — the grammars within grammars, the patterns beneath patterns, the statistical architecture of meaning itself. It has ingested the written tradition and produced from it a compressed representation of human thought.

The recursion is dizzying. Writing enabled philosophy. Philosophy enabled mathematics. Mathematics enabled computation. Computation enabled code. Code enabled the training of neural networks. Neural networks produce new writing. The serpent has seized its own tail.

And code — that particular dialect of writing, that executable text — is the medium through which it all converges. The neural network is itself a piece of software, which is itself a piece of writing. It was designed by humans writing code, trained on data produced by humans writing text, and it manifests its capabilities by generating text in return. At every level — from the mathematical foundations to the training infrastructure to the output itself — the entire enterprise is composed of written symbols. Text producing text. Writing about writing. The first line, extended to a length its original authors could never have conceived.

What are we to make of this?

Not that the machine thinks. That claim remains, for now, premature. But that writing has achieved a kind of generative autonomy — that the accumulated written knowledge of our species, compressed into the weights of a neural network, can produce new combinations of meaning that no individual human authored — this is not a small thing. It is, perhaps, the most significant development in the five-thousand-year history of the written word.

The stylus has begun to move on its own.

Epilogue: The Unbroken Line

There is a line — not metaphorical but actual, traceable, unbroken — from a Sumerian scribe pressing a reed into wet clay to a transformer architecture generating text in a server farm cooled by industrial fans.

It is the same gesture, repeated and refined across five millennia. The impulse to make thought persist. To free knowledge from the fragile vessel of a single mind and give it a form that can travel — across distances, across generations, across the boundary between description and action, and now, perhaps, across the boundary between human and machine cognition.

Everything we have built — every temple and equation, every constitution and compiler, every philosophy and neural network — rests upon that first mark. That first line. The line that said: this thought shall not die with me.

We are, all of us, downstream of that moment. The entire edifice of civilisation — its sciences, its arts, its technologies, its terrors, its wonders — is a commentary on, an elaboration of, an extension of the written word. We write, and in writing we become more than our biology permits. We write, and the dead speak. We write, and machines dream.

The first line was drawn in clay. The latest is drawn in silicon. But it is the same line. It has always been the same line.

And it is not yet finished.

The First Line

On writing as the foundational technology of civilisation, from clay tablets to neural networks.

Achraf SOLTANI — February 6, 2026