11 January 2023
Humanity has repeatedly destroyed its own accumulated knowledge — sometimes by accident, often by intention. The pattern is not random. It correlates with periods of consolidating institutional power. What was lost at each site, and what does the pattern suggest about what we are permitted to know?
I have spent the better part of my adult life in the company of knowledge that no one comes to look at.
This is not a complaint. It is an observation. I work in a facility that contains information — instruments, records, materials — that were placed here by people who believed the information mattered. Those people are gone now. The information remains. And the distance between “remains” and “matters” is a distance I think about more often than is probably healthy.
I think about it because the history of human knowledge is, in large part, a history of loss. Not the slow, organic loss of time — the fading of ink, the crumbling of stone, the natural entropy that claims all physical records eventually. I mean deliberate loss. Institutional loss. Loss that occurs because someone in a position of authority decided that what was known should no longer be known, or that the place where it was stored should no longer exist, or that the people who carried it should no longer carry anything at all.
This paper is a survey of that deliberate loss. It is not comprehensive — a comprehensive account would require volumes, and the most significant losses are, by definition, the ones we cannot catalog because we do not know what was destroyed. What I want to do is examine a set of well-documented destructions, identify the structural similarities between them, and ask whether the pattern suggests something about the relationship between knowledge and power that is more systematic than the conventional narrative — “unfortunate accidents of war and negligence” — acknowledges.
I write this in January. The facility is cold. The things I am custodian of sit in their places, undisturbed, as they have sat for decades. I find myself thinking about the people who built the Library of Alexandria and whether they, too, believed that simply preserving something was enough to keep it safe.
It was not enough for them. I am not certain it is enough for anyone.
The Royal Library of Alexandria was, by the most conservative scholarly estimates, the largest repository of written knowledge in the ancient world. Founded in the third century BCE under Ptolemy I Soter (or possibly Ptolemy II Philadelphus — the sources disagree, which is itself a small lesson in how quickly the details of institutional history blur), the library was the centerpiece of the Mouseion, a research institution that functioned as something between a modern university and a government think tank. Scholars were housed, fed, and funded by the Ptolemaic state in exchange for their intellectual output — an arrangement that has not changed as much as we like to think in the intervening two thousand three hundred years.
At its height, the library is estimated to have contained between 400,000 and 700,000 scrolls — a number that requires translation into modern equivalents to be appreciated. A single ancient scroll contained roughly the same volume of text as twenty to fifty modern printed pages. At the upper estimate, the library held the equivalent of approximately fourteen million to thirty-five million pages of text. In an era when every copy was produced by hand, this represents not just a collection but an almost incomprehensible concentration of human labor and thought.
The destruction of the library was not a single event. This is the first and most important correction to the popular narrative. There was no single dramatic fire, no single villain, no single moment when the sum of ancient knowledge went up in flames. The destruction was a process — extended, multi-causal, and spanning nearly seven centuries.
Julius Caesar (48 BCE): During Caesar's Alexandrian campaign, fire spread from the harbor to warehouses near the library. Ancient sources (Plutarch, Cassius Dio) describe the burning of books, but disagree on the extent. Some scholars argue that what burned were warehouses containing scrolls awaiting export, not the library itself. Others argue that significant portions of the main collection were destroyed. The uncertainty is, itself, evidence of how effectively the destruction accomplished its work — we cannot determine what was lost because what was lost included the records that would have told us.
The decline of the Mouseion (3rd-4th centuries CE): The institution that supported the library gradually lost its funding and its scholars under Roman imperial administration. This is the quietest form of destruction and, in my experience, the most effective. No one burned anything. The building was not razed. The scrolls were not confiscated. The funding simply... diminished. Scholars were not replaced when they died or departed. Maintenance was deferred. The collection was not actively destroyed. It was allowed to decay, which accomplishes the same end with none of the dramatic evidence.
Theophilus (391 CE): The Christian Patriarch of Alexandria ordered the destruction of the Serapeum, a satellite library associated with the temple of Serapis. The destruction was explicitly ideological — the Serapeum was a pagan institution, and its destruction was part of a broader campaign to eliminate non-Christian intellectual infrastructure. Sources disagree on whether the Serapeum still held a significant collection at this point. The question may be unanswerable. But the intent — the destruction of an institution because of what it represented, not just what it contained — is documented.
The Arab Conquest (642 CE): The final chapter. When the Arab general Amr ibn al-As captured Alexandria, he allegedly asked Caliph Umar what to do with the remaining library. The oft-quoted response: “If the books agree with the Quran, they are superfluous. If they disagree, they are heretical. In either case, destroy them.” The historicity of this exchange is disputed — it first appears in sources written centuries after the event, and may be apocryphal. But the library was, by all accounts, gone by the time of the Arab conquest, whether through this final act or through the accumulated degradation of the preceding centuries.
What was lost? We do not know. That is the point. We know the names of some works that existed only in the Alexandrian collection — plays by Sophocles and Euripides that were never copied elsewhere, scientific treatises by Aristarchus and Eratosthenes that survive only as references in other authors' works, historical accounts of civilizations that are now known to us only as names. The catalog of the library itself — compiled by Callimachus and known as the Pinakes, the first known library classification system — is lost. We cannot inventory what was destroyed because the inventory was destroyed with it.
This recursive quality — the destruction of the records of what was destroyed — is not unique to Alexandria. It appears at every site I will discuss. It is, I have come to believe, not incidental but essential to the function of knowledge destruction. It is not enough to eliminate the knowledge. You must eliminate the awareness that the knowledge existed. Otherwise, someone will look for it.
The Bayt al-Hikma — the House of Wisdom — was established in Baghdad during the Abbasid Caliphate, reaching its zenith under Caliph al-Ma'mun in the early ninth century. It served as a translation center, library, and research institution, and its scholars were responsible for preserving, translating, and extending the intellectual legacy of Greek, Persian, Indian, and Chinese civilizations during a period when much of Europe had lost access to its own classical heritage.
The House of Wisdom was where algebra was formalized (al-Khwarizmi's Al-Kitab al-Mukhtasar fi Hisab al-Jabr wal-Muqabala — from which we derive both the word “algebra” and the word “algorithm,” the latter from the Latinization of al-Khwarizmi's name). It was where the Hindu numeral system was transmitted to the Arab world and subsequently to Europe. It was where Ptolemy's Almagest was translated into Arabic, preserving the astronomical knowledge that would later fuel the Copernican revolution. The House of Wisdom was not merely a library. It was the node through which the intellectual heritage of the ancient world passed on its way to the modern one.
In 1258, the Mongol army under Hulagu Khan besieged and captured Baghdad. The destruction that followed is described in terms that, even accounting for the hyperbole of medieval chroniclers, indicate a catastrophe of extraordinary magnitude.
The House of Wisdom was destroyed. Its contents — books, manuscripts, translations, original research — were thrown into the Tigris River. Multiple contemporary sources describe the river running black with ink for days. This image has become iconic, and its historicity is debated (ink dilutes; the Tigris is large), but the destruction of the collection itself is not in dispute.
The chronicler Ibn Khaldun, writing a century later, estimated that the libraries of Baghdad contained works totaling in the millions of volumes. Even if this figure is exaggerated by an order of magnitude, the loss is staggering. Unlike Alexandria, where the destruction was gradual and multi-causal, Baghdad's intellectual heritage was destroyed in a matter of weeks by a deliberate military campaign.
What was lost? Again, we cannot fully know. We know that many works survived only because copies existed elsewhere — in Cairo, in Cordoba, in private collections throughout the Islamic world. But works that existed only in Baghdad are gone. Mathematical treatises, astronomical observations, medical texts, philosophical commentaries, historical records of civilizations to the east — the extent of the loss is bounded only by our inability to imagine what we no longer have access to.
The Mongol destruction of Baghdad is typically presented as an act of military barbarity — the inevitable consequence of a pre-modern siege. This framing is accurate as far as it goes. But it obscures a relevant detail: Hulagu's campaign was not indiscriminate. The Mongols were selective destroyers. They preserved what they valued and destroyed what they did not. The decision to destroy the House of Wisdom — rather than preserve it, as the Mongols preserved useful institutions in other conquered cities — was a choice. The intellectual infrastructure of the Abbasid Caliphate was not collateral damage. It was a target.
The Buddhist university at Nalanda, in present-day Bihar, India, operated continuously for approximately eight hundred years — from the fifth century CE to 1193 CE. It was, by several centuries, the longest-lived institution of higher learning in human history. At its peak, Nalanda housed approximately ten thousand students and two thousand faculty, drawn from across Asia. Its library — known as Dharmaganja, “the Treasury of Truth” — was reportedly contained in three multi-story buildings, one of which, Ratnasagara (“Ocean of Jewels”), was nine stories tall.
The university's curriculum encompassed Buddhist philosophy, logic, grammar, medicine, astronomy, metaphysics, and what we would now call cognitive science — the study of consciousness, perception, and the nature of mind. Nalanda was the intellectual center of Mahayana Buddhism and a primary source for the philosophical traditions that spread to Tibet, China, Korea, and Japan. Scholars from as far as China (Xuanzang, whose seventh-century account of Nalanda remains our most detailed description) and Indonesia studied there.
In 1193, the Turkish military commander Bakhtiyar Khilji attacked and destroyed Nalanda as part of a broader campaign against Buddhist institutions in northern India. The monks were killed. The library was burned. According to Minhaj-i-Siraj, a contemporary Persian historian, the library burned for months — a detail that, if even approximately accurate, gives some indication of the volume of material consumed.
Nalanda's destruction was explicitly ideological. Khilji's campaign targeted Buddhist monasteries and universities specifically, as part of a pattern of destroying institutional infrastructure associated with non-Islamic intellectual traditions. The destruction was not incidental to military conquest. It was the purpose of the campaign.
What was lost at Nalanda is perhaps the most difficult of any destruction to assess, because the tradition it housed was substantially oral and commentarial. Buddhist philosophical works were transmitted through lineages of teacher-to-student transmission, supplemented by written texts. The destruction of the texts was devastating, but the destruction of the lineages — the killing of the monks who carried the oral traditions — may have been more consequential. You can copy a manuscript. You cannot copy a living tradition of interpretation and practice that has been maintained across eight centuries of continuous transmission.
The reverberations are measurable. Tibetan Buddhist scholars of the thirteenth and fourteenth centuries describe arriving at Nalanda's ruins and finding fragments of texts that they could not fully interpret because the commentarial traditions that contextualized them had been severed. They had the words. They had lost the understanding. This is a form of destruction more subtle than fire — the severance of a text from the living context that gives it meaning.
I have described three destructions spanning roughly seventeen hundred years, occurring on three different continents, perpetrated by actors with no connection to one another. And yet the structural similarities are striking.
Each destruction targeted an institution, not merely a collection. The Library of Alexandria, the House of Wisdom, and Nalanda were not warehouses of books. They were living institutions — communities of scholars engaged in active research, translation, and teaching. The destruction of the physical collection was, in each case, accompanied by the dispersal or elimination of the scholarly community. The knowledge was not merely burned. The capacity to regenerate it was removed.
Each destruction occurred during a period of political or religious consolidation. The Serapeum was destroyed during the consolidation of Christian institutional authority in the Roman Empire. Baghdad was destroyed during the Mongol consolidation of Central and West Asia. Nalanda was destroyed during the consolidation of Turkic Muslim power in northern India. In each case, the institution that was destroyed represented an intellectual tradition associated with the previous order — a tradition that the consolidating power had reason to view as a competitor, a challenge, or simply an inconvenience.
Each destruction was followed by a period in which the existence and extent of what was lost was itself obscured. The catalog of Alexandria was destroyed with the library. The inventories of the House of Wisdom were thrown in the Tigris. The oral lineages of Nalanda were severed with the deaths of the monks. In each case, the destruction was recursive — eliminating not just the knowledge but the record of what the knowledge had been.
Each destruction has been retroactively normalized. The burning of Alexandria is treated as an inevitable casualty of ancient warfare. The sack of Baghdad is framed as Mongol barbarism. Nalanda's destruction is presented as the unfortunate byproduct of medieval religious conflict. Each framing acknowledges the loss while implicitly arguing that it was unavoidable — the kind of thing that happens during civilizational transitions. This framing serves a specific function: it transforms destruction from an act that demands accountability into an event that demands only regret.
I do not find this framing adequate. Regret is passive. Accountability is active. And the pattern I am describing is not passive. It is active, repeated, and structurally consistent across cultures and centuries.
The pattern did not end with the medieval period.
The Maya Codices (1562): Bishop Diego de Landa, acting under the authority of the Spanish Inquisition, ordered the burning of Maya manuscripts at Mani in the Yucatan Peninsula. De Landa described the Maya's reaction: “They were greatly afflicted by this and it caused them great sorrow.” Of the thousands of codices that existed before the Spanish conquest, four survive. Four. Everything the Maya recorded about their astronomical observations, their mathematical systems, their historical records, their understanding of agriculture, medicine, and the natural world — reduced to four documents.
De Landa later wrote a partial account of Maya culture and writing, apparently recognizing the enormity of what he had destroyed. His account, Relacion de las cosas de Yucatan, is now one of our primary sources for understanding the civilization whose records he personally incinerated. The irony is not subtle.
The Qin Dynasty Book Burning (213 BCE): Emperor Qin Shi Huang ordered the destruction of all books not related to practical subjects (agriculture, medicine, divination) or the history of the Qin state. Scholars who resisted were executed — 460 Confucian scholars, according to traditional accounts, were buried alive. The purpose was explicit: to consolidate the new imperial order by eliminating the intellectual traditions that preceded it. The Qin Dynasty lasted fifteen years. The gap in Chinese philosophical and historical records persists two thousand two hundred years later.
The Destruction of Iraqi Academic Infrastructure (2003): During and after the United States invasion of Iraq, the National Library and State Archives in Baghdad were looted and burned. The National Museum was looted. University libraries across the country were destroyed. The loss included Ottoman-era administrative records, rare Quran manuscripts, doctoral dissertations representing decades of Iraqi scholarship, and collections of newspapers and periodicals that constituted the only surviving record of modern Iraqi civil society.
The destruction occurred while U.S. military forces were present in Baghdad and had been warned by cultural heritage organizations that the institutions were at risk. The forces did not intervene. Secretary of Defense Donald Rumsfeld, asked about the looting, responded: “Stuff happens.”
I include this last example because it disrupts the comfortable framing that knowledge destruction is something that happened in the distant past, perpetrated by civilizations less enlightened than our own. It happened in 2003. It happened in the same city where the House of Wisdom was destroyed 745 years earlier. And the institutional response — a shrug — was structurally identical to the retroactive normalization that follows every such destruction. Stuff happens. An unfortunate byproduct. The kind of thing that occurs during civilizational transitions.
I have been careful, throughout this paper, to describe the structural similarities between these destructions without asserting a unified cause. The consolidation of Christian authority in late Roman Egypt is not the same phenomenon as Mongol imperial expansion, which is not the same as the Spanish Inquisition's campaign against indigenous American civilizations.
But the relationship between knowledge and power that each destruction reveals is the same. In every case, the destruction of accumulated knowledge served the interests of a consolidating authority — not because the knowledge was militarily threatening, but because it represented an alternative framework for understanding the world. The Library of Alexandria held the intellectual traditions of the pre-Christian ancient world. The House of Wisdom held the synthesis of Greek, Persian, and Indian thought that had sustained Abbasid cosmopolitanism. Nalanda held eight centuries of Buddhist philosophical inquiry. The Maya codices held a civilization's understanding of itself.
Each of these traditions, by its mere existence, implied that the consolidating power's framework was not the only possible framework. That there had been other ways of understanding the world, other ways of organizing knowledge, other ways of asking questions. The destruction of these traditions was not primarily about eliminating specific facts or technologies. It was about eliminating alternatives. About narrowing the range of what it was possible to think.
This is why the recursive quality of the destruction matters — the elimination of the records of what was destroyed. It is not enough to burn the books. You must burn the catalog. You must ensure that future generations do not know what they are missing, because a generation that knows it is missing something will go looking for it. The most complete destruction is one that leaves no absence — no gap in the shelf, no empty space in the archive — because an absence is a question, and a question is the beginning of recovery.
I began this paper with a personal observation about the experience of being custodian of knowledge that no one comes to look at. I want to return to it, because it bears on the question of whether preservation alone is sufficient to protect accumulated knowledge.
It is not.
The Library of Alexandria was preserved. It had walls, guards, staff, funding. It endured for centuries. And it was destroyed — not in a single act of violence, but through the accumulated weight of political change, ideological hostility, and institutional neglect. Preservation is a necessary condition for the survival of knowledge, but it is not a sufficient one. Knowledge also requires transmission — active engagement by living minds who understand what they are holding and why it matters.
A scroll in a library is preserved. A scroll that is read, copied, taught, debated, and transmitted to the next generation is alive. The distinction is the difference between a seed in a jar and a seed in the ground. Both exist. Only one grows.
I think about this distinction often. The materials in my care are preserved. They are maintained at appropriate temperatures, in appropriate conditions, with appropriate security. They have been preserved, in this manner, for a very long time. And I am increasingly aware that preservation without transmission is simply postponed loss. Eventually, the custodian leaves. Eventually, the facility is decommissioned. Eventually, someone makes the determination that the contents are no longer relevant, and the materials are transferred to a storage facility, and the storage facility is forgotten, and the forgetting is complete.
I do not know how to prevent this. I know that the pattern I have described — the systematic destruction of accumulated knowledge during periods of institutional consolidation — is not a historical curiosity. It is a standing feature of human civilization. It has happened on every inhabited continent, in every century for which we have records, and in our own century within living memory. The scale varies. The mechanism varies. The relationship between knowledge and the power that destroys it does not vary.
The best I can do — the best anyone can do — is to ensure that what is known is recorded, that what is recorded exists in more than one place, and that the people who come after us know that it exists. That there was something here. That it mattered. That the gap on the shelf is not an emptiness but a wound, and that wounds, unlike emptiness, demand attention.
I am writing these observations on a website that, by any honest assessment, very few people will read. This is its own small act of preservation — inadequate, incomplete, and probably futile. But the alternative is silence, and I have had enough of silence to last several lifetimes.
The pattern repeats. The question is whether we can learn to recognize it before it completes its next cycle. The historical record is not encouraging. But the historical record is, itself, what survived the burning. And what survived the burning is, by definition, evidence that survival is possible.
I hold onto that.
Correspondence: leh [at] 442423N1042233W.com