The upgrade that made the deposit cumulative at all, and the medium in which most of it is now carried. It evolved once, and the single root has consequences.

The eye evolved independently at least forty times. From the photoreceptor patches of flatworms to the compound eyes of insects to the camera eyes of vertebrates and cephalopods, natural selection found the same solution to the problem of light over and over, from different starting points, by different genetic routes, arriving at organs that do the same job through entirely different anatomies. It is one of the clearest demonstrations in biology that some solutions are so useful, and the problem so fundamental, that evolution will find them repeatedly given time.

Language is different. The evidence is that the biological capacity for language, in the specific sense of the compositional, recursive, symbolically referential system that upgrades f(·) and makes cumulative accumulation possible, evolved once. Or at least survives in a single line. All of the world's languages, from the click languages of the Kalahari to Mandarin to Basque to Swahili, appear to descend from one ancestral population: a proto-language community, almost certainly in Africa, somewhere between 50,000 and 100,000 years ago, that first crossed the threshold into full modern language. It may have crossed more than once. If so, the competition between the early linguistic populations was as terminal as any in the book, and one deposit outlasted the rest. The surviving record has one root either way. Vocabulary and grammar diverge too fast for direct comparison beyond a few thousand years, but the deep features (recursion, compositionality, displacement in time and space) are universal in a way that implies common descent rather than independent invention. The reason may be a small step change in the integrand that introduced a complex axis, as related in the Calculus of Belief in Part Three.

This matters for the civilisational record. The eye's forty inventions meant that vision spread across the animal kingdom from many sources. The single invention of language meant that the civilisational deposit has one cognitive root and many language-family branches, its diversifications the bifurcations produced by migration, isolation and time. Every written record, every philosophical tradition, every legal system, every scientific corpus and every poem traces back to a single population that crossed a threshold the rest of the animal kingdom has not crossed since, and that our own lineage crossed, apparently, only once.

What followed that crossing is the subject of everything else in the book. The particular story here is language as a technology: the medium through which civilisation transmits, extends, competes and sometimes dies.

The first language communities left no written record. What we know comes from the comparative linguistics of reconstructed proto-languages, the genetic record of population movements and the archaeology of the periods to which expansion can be roughly dated. The great language families (Indo-European, Sino-Tibetan, Afro-Asiatic, Austronesian, Niger-Congo) are each a story of expansion: a population of speakers whose deposit proved resonant enough and competitive enough to spread across enormous ranges, carrying its language with it or displacing the languages it met. Sanskrit, among the oldest attested of the Indo-European line, was the medium through which the philosophical, mathematical and literary deposit of Indian civilisation travelled.

Chinese is a different solution to the same problem, with consequences that the Western scholarly tradition has not fully taken in. The writing system, developed across the second millennium BCE and stabilised under the Han, is logographic rather than alphabetic. It encodes meaning where an alphabet encodes sound. That gave it a property no other major writing system has: mutual intelligibility across mutually unintelligible speech. A Cantonese speaker and a Mandarin speaker, unable to follow each other's talk, can read each other's text. The script became the shared frame of a civilisation spread across a vast and linguistically diverse geography, and the imperial examination system, which required mastery of the classical corpus as the price of administrative participation, became the most systematic and durable educational transmission pathway ever built: two thousand years of standardised examination in one body of text, a civilisation's formative process institutionalised and applied at scale.

Arabic reached its civilisational moment through a specific conjunction. The Quran bound the language to a sacred form of extraordinary resonance and established a linguistic community that within a century ran from the Arabian Peninsula to the Atlantic coast of North Africa. But Arabic's most consequential contribution was not its own accumulation, substantial as that was. It was its function as the transmission medium for the Greek corpus. The Abbasid translation movement carried the deposit of an entire civilisation into a new linguistic medium, where it was extended and enriched by scholars who brought knowledge of their own to the encounter. Ibn Sina's Canon is not Galen translated. It is Galen compounded with Arabic clinical and philosophical inheritance, which is what the Integral forecasts when two rich traditions achieve resonance rather than coexistence.

The Renaissance rediscovered the same process. The Greek texts that reached fifteenth-century Florence (many through Arabic transmission, some with Byzantine refugees fleeing the Ottoman advance) did not arrive into empty ground. They arrived into parameter sets that centuries of Latin scholarship and Christian intellectual culture had built into a state capable of receiving them. The resonance was not inevitable. It took humanist scholars, Medici patronage and the printing press to realise it. When it occurred, it rebuilt European civilisation more deeply than any event since the fall of Rome. Latin had been Europe's shared written frame for a millennium, persisting above the divergence of spoken tongues. The Renaissance added Greek to that substrate and doubled the accessible portion of the classical deposit, providing the empirical and philosophical foundation on which the Scientific Revolution built.

The press extended the bandwidth of this recovery at a rate no previous transmission technology had approached. What had required the sustained labour of monastic copyists was now propagated in editions of hundreds, then thousands, then tens of thousands. Luther's theses did not merely spread faster than Wycliffe's. They spread too fast for institutional suppression to be possible: the first modern instance of an idea moving through a new medium more quickly than the existing power structure could contain it.

English reached its imperial moment through a conjunction that mirrors the Arabic case at larger scale. The British Empire, at its maximum extent in the early twentieth century, administered roughly a quarter of the world's land and a comparable fraction of its people, not through linguistic uniformity but by imposing English as the language of law, administration, education and commerce across an enormous diversity of existing communities. The consequences were not uniform. In Ireland, India and the Caribbean, English arrived as an instrument of subjugation, displacing vernacular traditions with deliberate thoroughness and installing a foreign frame as the condition of public life. In Australia it arrived alongside the near-destruction of the world's oldest continuous cultural traditions: more than two hundred distinct language communities, each the carrier of knowledge about a specific environment refined over 50,000 years, reduced within two centuries to a handful of survivors.

And English also did what every imperial language before it had done, and what Arabic did for Greek. It became the medium through which very different traditions could meet. The post-colonial literature of Chinua Achebe, V. S. Naipaul and Arundhati Roy, work that turned the imperial language against the imperial project, is the process running as the model predicts: a medium adopted under coercion becoming the vehicle for what its original carriers had not anticipated and could not control. The language transmits what the language carries, and what it carries is set by the integrals that use it, not by the power that imposed it.

What English has become in the early twenty-first century has no precedent. A global lingua franca, spoken as a first or second language by more than a billion and a half people, serving as the primary medium of science, commerce, diplomacy and digital communication at once. Latin united a few million people around the Mediterranean. Classical Chinese united a few hundred million across East Asia. English now mediates the interactions of a civilisation whose parameter sets are more diverse, and more rapidly interacting, than at any previous point. The consequence for the civilisational accumulation is that the outer frame, the shared medium that makes competition generative, now extends, imperfectly but substantially, across the whole human world for the first time.