1 What makes an event “historical”?

Most of what has ever happened leaves no trace. The small fragment that does — and that gets taken up, argued over, and passed on — is what we end up calling history. The difference between the two is not about importance in any obvious sense: billions of people have lived and died without leaving a single sentence in the historical record, while a relatively small number of political events, battles, and rulers have generated thousands of volumes. That disparity is not neutral. It is a choice — and choices have reasons.

1.1 Carr’s Dialogue

The Unknown Soldier names the pattern: the past is not handed to the historian — the historian selects a tiny, freighted piece of it and calls that selection history. E. H. Carr named the principle that makes the selection work.

E.H. Carr opens What is History? (1961) with a deceptively simple formula:

“My first answer therefore to the question ‘What is history?’ is that it is a continuous process of interaction between the historian and his facts, an unending dialogue between the present and the past.” — E. H. Carr, What is History?, Chapter 1²

The dialogue metaphor is doing real philosophical work here. A dialogue is not a one-way transmission. The present shapes which questions we ask of the past; the past shapes how we understand the present. This is not relativism — Carr is not saying that any account of the past is as good as any other. He is saying that there is no account of the past that was produced from outside any present. The historian is always located somewhere.

Carr’s provocation extends to the facts themselves. He distinguishes between historical facts (facts that historians have selected and judged significant enough to discuss) and the past as it was (the everything that happened). “The historian is necessarily selective,” he writes. “The belief in a hard core of historical facts existing objectively and independently of the interpretation of the historian is a preposterous fallacy.”³

This sounds radical. But consider: the assassination of Archduke Franz Ferdinand on 28 June 1914 is unambiguously a historical fact. The assassination, on the same day, of a Serbian farmer whose name we do not know, is also a fact about the past — but no one calls it a historical fact. Not because it didn’t happen, but because no historian has made it the subject of a claim. The “factness” of a historical fact is already the result of a judgement about significance.

1.2 Von Ranke and the Dream of Objectivity

Carr’s nemesis (and the villain of his first chapter) is Leopold von Ranke, the 19th-century German historian who articulated the ideal of showing the past wie es eigentlich gewesen — as it actually was.⁴ Von Ranke wanted history freed from the moralising, the nationalist mythology, and the speculative philosophy that had characterised earlier historiography. He wanted evidence: documents, archives, primary sources. History should be grounded in what actually happened, not in what we wish had happened.

This was a genuine and important reform. The problem is the phrase “as it actually was.” Ranke used it to describe a goal. But the goal is unreachable. Any account of the past involves selection, arrangement, and interpretation. Even a chronicle — a simple list of dated events — reflects choices about what counts as an event worth listing. There is no record of everything; there is no viewpoint from nowhere; there is no neutral vocabulary.

The myth of von Rankean objectivity is not that the past cannot be known. It is the myth that the historian can disappear from the account — that the past can speak for itself. It cannot. It speaks only through the questions historians ask.

A note on Ranke himself: the figure Carr attacks is part real, part caricature. Specialist historians of historiography (Iggers, Krieger, Novick) point out that eigentlich in nineteenth-century German is closer to “essentially” or “in its own terms” than to “actually” in the modern English sense, and that Ranke was a Lutheran idealist who believed history could trace a divine purpose, not the naive positivist for whom facts speak for themselves. Whether the revisionist reading rehabilitates Ranke or whether Carr’s target was real enough that the rehabilitation overstates the case, is itself a continuing dispute in the field.⁵

1.3 Earlier Theorists of History: Ibn Khaldun and Sima Qian

Carr and Ranke are nineteenth- and twentieth-century European disputants in a much older argument. Two earlier non-European traditions theorised history with comparable rigour, and a course that takes its categories only from the European debate misses what those traditions saw.

The Tunisian-Andalusian scholar Ibn Khaldun, in the Muqaddimah (1377), is widely credited as the founder of historiography as a methodological discipline. His central concept, ʿaṣabiyyah — usually translated “social cohesion” or “group feeling” — is offered as the underlying explanatory variable in the rise and fall of dynasties: a tribal coalition with strong ʿaṣabiyyah conquers a sedentary city; over three or four generations, urban life corrodes its cohesion; a new coalition with stronger ʿaṣabiyyah succeeds it.⁶ What is striking for the present discussion is not the cyclic theory itself (which is contested) but the methodological apparatus Ibn Khaldun proposes for evaluating historical claims: the Muqaddimah opens with an explicit catalogue of the errors historians make — partisanship, credulity, ignorance of context, mistaking the impossible for the merely improbable, exaggeration of numbers — and a proposed corrective: every historical report must be checked against the structural conditions of the society in which it is alleged to have occurred. The procedure is a four-century anticipation of the source-criticism that will later be associated with Niebuhr and Ranke.

The Han Chinese historian Sima Qian (司馬遷, c. 145–86 BCE), in the Shiji (史記, Records of the Grand Historian), develops a different methodological apparatus. The Shiji organises history into five formats — benji (basic annals of rulers), biao (chronological tables), shu (treatises on institutions), shijia (hereditary houses), liezhuan (biographies of representative figures) — that together index the same events from multiple structural angles. Where Ibn Khaldun’s strength is causal analysis, Sima Qian’s is the explicit refusal of any single narrative line: by writing the same history five times in five formats, the Shiji shows that an event looks different from the perspective of dynastic legitimacy than from the perspective of institutional development than from the perspective of an individual life.⁷ Both traditions are not “anticipations” of European historiography; they are full-developed alternatives to it. Their absence from a course on the philosophy of history is itself an act of historiographical selection.

1.4 Selection and Significance

Why do some events become historical? Several mechanisms are in play:

Survival of evidence. We know what we have records of. Records survive unevenly: royal decrees, religious texts, military campaigns, official correspondence — these were stored, copied, and archived. The interior life of ordinary people, informal agreements, oral cultures — these mostly vanish. The archive is already a filter.

Institutional priorities. Historical scholarship is produced by universities, publishers, and archives — institutions with their own histories of who they admitted, what they funded, and what they considered worth studying. These priorities have systematically favoured certain kinds of history over others.

Narrative shape. Some events lend themselves to narrative — they have a clear beginning, middle, and end; clear actors; dramatic turning points. Others are too diffuse, too long, too unmarked by crisis to tell as a story. The long, slow deterioration of soil fertility in the Roman provinces is historically significant, but it is harder to narrate than the Battle of Actium.

Present concerns. What we want to know about the past is partly determined by what matters to us now. The explosion of scholarly attention to the history of gender, race, class, and sexuality since the 1970s was not driven by the discovery of new evidence. It was driven by political and social movements that changed what questions seemed important.

1.5 Questions to Argue About

• If historical facts are selected rather than given, does that mean history is just fiction? What is the difference between the two?
• Is Carr right that the historian’s present always shapes the questions they ask? If so, is this a problem to be solved, or a feature to be acknowledged and worked with?
• Von Ranke’s ideal of showing history “as it actually was” has been criticised as naive. But is the alternative — accepting that history is always perspectival — a dangerous concession to relativism and manipulation?
• The daily life of a Roman slave is not in the historical record. Is that an epistemological problem (we can’t know it) or an ethical problem (we chose not to care)? Or both?

2 Can we be objective about the past?

The question seems almost self-answering, and the history of historiography is littered with historians who thought so — and were wrong in exactly opposite directions. Half said: of course objectivity is possible; apply the methods, follow the evidence, and the truth will emerge. The other half said: of course it is impossible; you are a person, in a time and place, with interests and assumptions, writing about people and events that are no longer accessible to you directly. Both answers are too quick. “Objectivity is impossible” can mean several very different things, and not all of them are true.

2.1 The Historikerstreit (1986–87) and the Catechism Debate (2020–22)

In 1986, the German historian Ernst Nolte published a newspaper essay suggesting that the Holocaust, while not to be minimised, was in part a response to the prior terror of Stalinist collectivisation, and that Auschwitz should be understood in the context of other twentieth-century mass atrocities.¹² The essay triggered a prolonged controversy among German historians — the Historikerstreit, or historians’ dispute — in which Jürgen Habermas accused Nolte of attempting to “relativise” the Holocaust for nationalist political purposes.¹³ The debate was about whether comparative history was a legitimate method when applied to an event of unique moral weight, and about whether a historian’s political location determines the meaning of their historical claims. Nolte insisted he was engaging in objective comparative analysis; Habermas argued that no such analysis was politically innocent given the specific context of 1980s West Germany. Karl Mannheim’s relationism — the view that all historical knowledge is perspectival but not therefore arbitrary — was being tested in real time.¹⁴

The same dispute was relitigated three decades on. In April 2020, Germany’s Federal Commissioner against antisemitism, Felix Klein, called for the cancellation of an invitation to the Cameroonian philosopher Achille Mbembe to address the Ruhrtriennale festival, on the grounds that Mbembe’s parallels between Israeli policy and apartheid South Africa constituted Holocaust relativisation.¹⁵ From May 2021, Die Zeit published a sequence of essays under the heading Katechismus-Streit; A. Dirk Moses’s “The German Catechism” argued that comparative historians of the Holocaust were being silenced by a German civic catechism (Holocaust unique, comparison is relativisation).¹⁶ Friedländer, Zimmerer, Assmann, Diner, Neiman, and Rothberg responded across the German press; the dispute ran through 2022 without resolution.¹⁷ What the Historikerstreit, the Catechism Debate, and the Bergier Commission (info box above) all establish — in different national contexts, decades apart — is that the question of historical objectivity cannot be answered in the abstract because the stakes of any particular historical claim are always embedded in a present-day political situation. Switzerland’s reckoning came under external pressure (Eizenstat, WJC); Germany’s came from inside (Nolte, then Mbembe); the methodological question they share is the same.

2.2 Collingwood’s Re-enactment

R.G. Collingwood, in The Idea of History (1946), argues that historical knowledge is fundamentally different from scientific knowledge.¹⁸ In science, you study objects — things that exist independently of human minds. In history, you study actions — and actions are not mere physical events. They have an inside as well as an outside.

The outside of Julius Caesar’s crossing of the Rubicon in 49 BCE is a physical event: a man with an army moving across a river. The inside is Caesar’s decision, his reasons, his understanding of the situation. To understand the historical event is to understand the inside — and to understand the inside, the historian must re-enact Caesar’s thought in their own mind. Not imagine what Caesar was feeling, but reconstruct what he was thinking: what arguments, what evidence, what considerations were available to him and led to his choice.¹⁹

“All history is the history of thought.” — R.G. Collingwood, The Idea of History, Part V²⁰

Collingwood is making two claims at once: about what history is, and about how it is known. History is not the study of past events but the study of past thinking. And it is known not by observation (you cannot observe Caesar’s thoughts) but by a kind of disciplined sympathetic reconstruction.

2.3 Thucydides vs. Herodotus

The contrast between the two great ancient Greek historians is the founding dispute in the philosophy of history. Herodotus (c. 484–425 BCE), the so-called “father of history,” collected stories, traditions, and accounts from his travels.²¹ He included divine intervention, fabulous creatures, and hearsay — he often reported things he did not believe, on the grounds that his job was to record what people said, not to adjudicate truth. His history is rich, capacious, and wonderfully human.

Thucydides (c. 460–400 BCE) was contemptuous of this approach. His account of the Peloponnesian War aims to be a “possession for all time,” not an entertainment for the moment. He strips out divine causality, exercises rigorous source criticism, and constructs causal explanations in terms of human rationality, power, and interest. He acknowledges that he has reconstructed the speeches in his history (“keeping as close as possible to the general sense of what was actually said”) — an extraordinary moment of methodological self-awareness.²²

Which approach is better? Thucydides is more rigorous, more usable, more scientific in aspiration. But Herodotus’s willingness to record without fully endorsing preserves more of the ancient world’s texture and its people’s self-understanding. Both are doing something real. The question is what history is for.

2.4 The Myth of the View from Nowhere

Objectivity, in its ideal form, would require a perspective from nowhere — a standpoint outside all historical, cultural, and personal situatedness. This is the “God’s eye view,” the view from which all particular perspectives cancel out and reality is seen as it truly is.

The philosopher Thomas Nagel called this the “view from nowhere” in his 1986 book of the same name — and his point was not that it is desirable but that it is unreachable.²³ For human knowers, there is always a somewhere. You are always looking from a position.

But here is the distinction that matters: acknowledging that you have a perspective is not the same as saying all perspectives are equal. A German historian writing about the First World War in 1930 does not have the same perspective as a French historian writing about it in 1950. Their assumptions, their cultural memories, their prior commitments differ. These differences affect what they see and how they interpret it. The question is whether those differences are acknowledged and disciplined, or whether they operate invisibly.

One position holds that acknowledged perspectivalism is not a concession to relativism but a methodological tool: knowing your position — what interests you might have, what stories you have been told, what you might want to find — allows you to be on guard against those distortions. The opposing position holds that this self-knowledge is at best therapeutic and at worst a licence for displaying one’s perspective rather than disciplining it. The defeasible empirical question — whether historians who name their perspective in fact produce less biased work than those who do not — is not settled.

2.5 Questions to Argue About

• Collingwood says history is the history of thought, and the historian re-enacts past thinking. But can we really re-enact the thoughts of people from radically different cultures and times? What are the limits?
• Is the contrast between Thucydides and Herodotus really a contrast between rigour and richness? Or is Herodotus’s approach capturing something that Thucydides’s rigour necessarily excludes?
• Can a historian ever be more objective by acknowledging their perspective, rather than claiming to have none? What would that look like in practice?
• Is there a difference between a biased account and a perspectival account? If so, what is it?

3 How do historians actually know?

The historian cannot visit the past. They cannot interview Caesar or photograph the Somme in 1916. They work with traces — documents, objects, images, memories — that the past has left behind. The epistemological question is how much we can infer from traces, and how confident we can be in those inferences.

3.1 Primary and Secondary Sources

The standard distinction: a primary source is a document or artefact from the time and place being studied — an eyewitness account, a contemporary newspaper, a government decree, a physical object. A secondary source is an interpretation produced later — a historian’s account, a textbook, a documentary.

The binary itself is largely a TOK-textbook frame; working historians use a more graded vocabulary. Marc Bloch, in Apologie pour l’histoire ou Métier d’historien (drafted 1941–1943, published posthumously 1949 as The Historian’s Craft), distinguished intentional witnesses (those who set out to inform posterity) from unintentional ones (parish registers, customs ledgers, court depositions, tax rolls), and argued that the unintentional sources are usually the more trustworthy because no one tried to shape them for the historian’s benefit.²⁸ Ginzburg’s evidential reading extends this: the documents that recovered Menocchio were Inquisition transcripts — involuntary witnesses produced by an institution with no interest in posterity. The contemporary digital-archive turn dissolves the binary further, since when a “primary source” is the result of a database query against a corpus assembled by curators, the boundary between document and interpretation has already been crossed before the historian opens the file.

The distinction is useful but not as clean as it looks. A primary source is not automatically more reliable than a secondary one. Several problems arise:

Eyewitness unreliability. Psychological research on memory (Elizabeth Loftus’s work on memory malleability is the landmark here — see Eyewitness Testimony, 1979, and Loftus and Palmer, “Reconstruction of Automobile Destruction,” Journal of Verbal Learning and Verbal Behavior 13, 1974) has demonstrated repeatedly that eyewitness accounts are not recordings.²⁹ Memory is reconstructive, not reproductive. What people remember is shaped by what they were told afterward, by their emotional state at the time, by subsequent information, and by the stories that make sense. Eyewitness testimony is highly persuasive and highly fallible simultaneously.

The wreck of the Titanic (1912) is a useful case: survivors gave contradictory accounts of whether the ship broke in two before sinking. The wreck, when found in 1985, showed it had broken in two.³⁰ Eyewitness testimony said otherwise — or rather, accounts divided, and the majority (in several oral history collections) said the ship sank intact. The physical evidence was necessary to settle what the eyewitnesses could not.

The problem of the author. Every primary source was produced by someone, for a purpose, with an intended audience. Julius Caesar’s Gallic Wars is a primary source for the Roman conquest of Gaul — and also a piece of political propaganda written to justify Caesar’s command and enhance his reputation in Rome.³¹ Both things are true simultaneously. The source has evidential value and must be read against the grain.

The problem of the archive. Not everything was recorded; not everything recorded was preserved; not everything preserved is accessible. Derrida, in Archive Fever (Mal d’Archive, 1995), argues that the archive is not a neutral repository of the past but a political institution: what gets archived, who has the right to deposit and access records, and what gets destroyed are all exercises of power.³³ The Haitian historian Michel-Rolph Trouillot disaggregates this problem into four distinct moments at which power produces silence: the making of sources, the making of archives, the making of narratives, and the making of history “in the final instance.”³⁴ Source-criticism that polices only the first moment leaves the other three untouched.

3.2 The Detective Model

Historical reasoning has been compared to detective work — and the comparison is illuminating. The logical structure is abductive reasoning, or inference to the best explanation: a term Gilbert Harman introduced in 1965, though the underlying idea goes back to C.S. Peirce.³⁵ W.H. Walsh described this reasoning in historical contexts;³⁶ Ginzburg, in “Clues: Roots of an Evidential Paradigm” (1979), developed it into a full account of the “conjectural” or “clues” paradigm of knowledge, tracing it from the tracking of animals, through medical diagnosis, to Sherlock Holmes and historical method.³⁷ The historian, like the detective, does not have direct access to what happened. They have evidence — traces, clues, inconsistencies — and must reason to the most probable account of what happened.

This model has real advantages. It is honest about the role of inference and imagination. It makes explicit that historical conclusions are probabilistic rather than certain. And it connects history to a broader tradition of abductive reasoning (inference to the best explanation) used across many domains.

The detective figure is closer to what Ginzburg called the evidential paradigm: reading symptoms — the involuntary trace, the inadvertent detail — the way Morelli read brushstrokes or a physician reads a body. Critics of the analogy press further: the historian does not only read clues left behind, the historian also assembles the case, decides what counts as a clue, and sets the stakes of the investigation. Hayden White’s Metahistory (1973) is the strongest version of this critique — the choice of plot structure (tragedy, romance, satire) is, on his account, already an interpretive act rather than a finding of the evidence.³⁸ Whether the constructive elements White names amount to a difference in degree from detective inference, or whether they make the detective analogy actively misleading, is itself disputed.

The case of Cleopatra VII of Egypt (69–30 BCE) illustrates the limits. Almost everything we know about her comes from sources written after her death, mostly by Roman authors, mostly hostile — she was the enemy they had defeated and the woman they wished to portray as a seductress and a threat. Plutarch’s account was written 120 years after her death.³⁹ We have no contemporary Egyptian sources that she commissioned or endorsed. We know what her enemies chose to say about her, filtered through 150 years of subsequent retelling. The detective work here involves asking: what would a hostile Roman source leave out? What does the pattern of silence tell us?

3.3 Synthetic Media and the 21st-Century Doctored Source

The Stalin photograph — Yezhov erased from the canal walk — is the canonical example of the manipulated historical source. The contemporary equivalent is of a different order: AI-generated synthetic media that can produce video and audio of historical figures saying things they never said. Forensic historians already encounter such material in online archives that will become the primary sources of future historians. The standard tools of source criticism — provenance, corroboration, internal consistency, comparison with independent sources — apply here, but they are no longer sufficient alone. Technical authentication (digital forensics, cryptographic signing) is becoming as important as traditional philology.

Forced Fork.Does this change the historian’s craft, or merely intensify Bloch’s old problem of the intentional witness?Position A: it changes the craft. The cost of fabrication has collapsed (in 1922 a forged document required a forger; in 2024 it required a prompt), the mechanism of transmission is opaque in a new way (a viral video with no chain of custody is not analogous to a forged letter with a paper trail), and the producer is a model whose outputs no individual is fully responsible for — a real change in the conditions of source criticism. Position B: it intensifies but does not change. Bloch’s distinction (The Historian’s Craft, Ch. II §2) between intentional and unintentional witnesses — the witness who means to convey a record vs. the trace left in passing — already covers the case. A deepfake is an intentional witness whose intentions were a fraud; the historian has always had to ask who made this artefact and why, and the techniques (provenance, corroboration, comparison with independents) extend to the new medium. Choose.

3.4 Questions to Argue About

• If primary sources are not automatically more reliable than secondary ones, what should historians prioritise? What makes a source trustworthy?
• The archive preserves what powerful institutions chose to keep. Does this mean history is inherently skewed toward the powerful? Or can historians compensate?
• Historical reasoning is described as inference to the best explanation. But what makes one explanation “best”? Is this an epistemic criterion, or does it smuggle in values?
• We know almost nothing about Cleopatra except through hostile sources. Should historians attempt a reconstruction, knowing it might be wrong, or acknowledge the limit and leave it blank?

4 Does the present shape how we see the past?

Of course it does. But the interesting question is exactly how, and whether this is always a distortion or whether it can be a source of insight.

4.1 Hindsight Bias

Knowing how things turned out makes them seem inevitable in retrospect. This is hindsight bias — one of the most replicated findings in cognitive psychology, first experimentally documented by Baruch Fischhoff in “Hindsight ≠ Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty” (Journal of Experimental Psychology: Human Perception and Performance 1.3, 1975) and later integrated into Kahneman’s broader account of cognitive biases (Thinking, Fast and Slow, 2011, Part III).⁴⁴⁴⁵ People who are told the outcome of an event consistently overestimate how predictable that outcome was. They construct a narrative that makes the outcome seem to flow naturally from the antecedent events, when in fact the outcome was deeply contingent.

The problem for history is that historians almost always know the outcomes. The historian writing about the July Crisis of 1914 knows that the crisis led to the First World War. This knowledge is almost impossible to bracket. Every decision by the Austrian government, every telegram from the Kaiser, every calculation by the Serbian Foreign Ministry gets read as a step toward war — because it was, in fact, a step toward war. But the participants didn’t know this. They were making decisions in radical uncertainty, in which many outcomes were genuinely open.

A historian who reads 1914 through the lens of its outcome misunderstands the choices made. The actors were not walking knowingly into catastrophe. They were miscalculating, bluffing, and misreading signals in real time.

The serious historiography of July 1914 has moved well beyond the general claim that “everyone has hindsight bias.” Fritz Fischer’s Griff nach der Weltmacht (1961) located the cause in deliberate German bids for continental hegemony and triggered the generation-long Fischer-Kontroverse; the controversy unfolded in a Cold War-era West Germany in which the question of national guilt was politically live, and several commentators on the controversy (Moses, Mombauer) have argued that the reception of Fischer cannot be cleanly separated from that context.⁴⁶ Christopher Clark’s The Sleepwalkers (2012), by contrast, redistributed responsibility across all the chancelleries of Europe.⁴⁷ The reading that its post-2008 reception was helped by an EU intellectual climate looking for a story of shared mismanagement is plausible but not established — Clark’s own argument turns on archival work in Vienna and Belgrade, and the book’s reception in Britain and France was at least as warm as in the EU institutions to which it is sometimes ascribed. The live methodological question is therefore narrower than “external politics drives historiography”: whether, in any given case, the historians’ moral and political commitments are demonstrably patterning the reading — and what evidence would settle that, beyond noting that the pattern is suggestive.

4.2 Revisionism and Its Uses

Historical revisionism is often treated as a term of abuse — Soviet propaganda revised history, so revisionism is manipulation. But as a scholarly practice, revisionism is simply the revision of previous historical accounts in light of new evidence or new questions. It is how history progresses.

Sheila Rowbotham’s Hidden from History (1973) is a case in point.⁴⁸ Rowbotham argued that conventional political and labour history had systematically ignored the experiences of women — not because the evidence didn’t exist, but because historians hadn’t looked for it and the institutional structures of the discipline had marginalised it. Her work, and the feminist historiography that followed, did not invent a past; it recovered one that had been overlooked.

The Russian Revolution is perhaps the most striking example of history as interpretive battleground. The same sequence of events — the collapse of Tsarist rule, the Provisional Government, the October events of 1917 — has been described as a “popular uprising” (by Soviet historians and by many Western sympathisers), a “Bolshevik coup” (by anti-communist historians), and a “complex, multi-actor power struggle in a context of total war” (by historians attempting to bracket ideological commitments).⁵⁰ Each description implies a different set of facts as relevant, a different causal story, and different contemporary lessons.

4.3 Orwell’s Warning

George Orwell understood something important about the political stakes of controlling historical memory:

“Who controls the past controls the future; who controls the present controls the past.” — George Orwell, Nineteen Eighty-Four (1949), Part Three, Chapter 2⁵¹

The novel’s Ministry of Truth (a masterpiece of Orwellian naming) employs Winston Smith to rewrite newspaper archives so that Party pronouncements are never contradicted by the historical record. When an ally becomes an enemy, all records of the alliance are destroyed. When a prediction fails, all records of the prediction are destroyed. The past is remade continuously to serve the needs of the present.

Orwell’s novel is fiction; the practice it describes is historical fact. The most notorious example is Stalin’s doctored photographs: as political rivals were executed or fell out of favour, they were literally removed from official photographs, so that the image of a parade or a meeting would show fewer people than had actually been there.⁵² The photograph, one of the most seemingly objective of primary sources, was a tool of active historical falsification.

4.4 Questions to Argue About

• Can a historian ever successfully bracket their knowledge of how events turned out? If not, is hindsight bias an unavoidable distortion, or can it be managed?
• Is all historical revisionism potentially legitimate, or are some narratives so well-established that challenging them requires exceptional evidence? Who decides which is which?
• Orwell’s dystopia is fictional, but the practices it describes (rewriting archives, removing people from photographs) have real historical instances. What mechanisms prevent this in democratic societies? Are they reliable?
• Can the present’s concerns improve our understanding of the past (by highlighting what was overlooked), or do they always distort it (by projecting contemporary categories backward)?

5 Is history made by individuals or structures?

Tolstoy devotes the second epilogue of War and Peace to this question — fifty pages of philosophy that a great many readers skip, which is a shame, because it is where the novel does its most serious thinking. He argues that Napoleon was not the cause of the 1812 campaign; he was a puppet of forces he neither created nor understood, performing a role that history had already written. Carlyle, who argued the opposite with equal force, would have found this intolerable. There are two visions of history’s motor. In one, history is driven by exceptional individuals — their genius, their will, their decisions. In the other, history is driven by impersonal forces — economic pressures, social structures, demographic shifts, geographic constraints — and individuals are swept along. Both visions capture something true, and both mislead if pushed too far.

5.1 The Great Man Theory

Thomas Carlyle stated the individualist thesis with characteristic force:

“The history of the world is but the biography of great men.” — Thomas Carlyle, On Heroes, Hero-Worship, and the Heroic in History (1841)⁶⁰

Carlyle meant this seriously. Napoleon is not an expression of larger historical forces; he is the cause of a new historical configuration. Without Napoleon, Austerlitz doesn’t happen, the Civil Code doesn’t happen, the redrawing of Europe doesn’t happen. The individual, at the decisive moment, makes history rather than merely enacting its logic.

The view has obvious appeal. It explains the felt importance of individual decisions. It coheres with our ordinary moral intuitions: we hold people responsible for what they do, which implies their actions matter. And it fits the biographical and narrative forms that dominate historical writing.

5.2 Marx’s Counter

Marx’s counter is not that individuals don’t matter, but that the conditions within which individuals act determine what is possible:

“Men make their own history, but they do not make it as they please; they do not make it under self-selected circumstances, but under circumstances existing already, given and transmitted from the past.” — Karl Marx, The Eighteenth Brumaire of Louis Bonaparte (1852), Chapter I⁶¹

For Marx, the real drivers of history are economic structures — the relations of production, the conflict between classes, the material interests that shape what people want and what they can imagine wanting. Napoleon couldn’t have done what he did if the conditions — the breakdown of the Old Regime, the revolutionary mobilisation of the bourgeoisie, the armies that the Revolution created — hadn’t existed. The conditions made someone like Napoleon possible and perhaps necessary; the individual is the vehicle for a process that has a logic of its own.

5.3 Tolstoy’s Critique

Leo Tolstoy, in the second epilogue to War and Peace (1869), pursues this argument with striking philosophical force. He is writing about Napoleon and the campaign of 1812, and his claim is that the attribution of historical causality to great individuals is a systematic illusion:⁶²

“Kings are the slaves of history. History, that is, the unconscious, swarmlike life of mankind, uses every moment of a king’s life as an instrument for its purposes.” — Leo Tolstoy, War and Peace, Book IX, Chapter 1, trans. Pevear & Volokhonsky⁶³

Tolstoy’s image is of a ship moving through water. The wake seems to be caused by the ship’s propeller — but the propeller is itself driven by the engine, which is driven by the coal, which was dug by miners, and so on. Cutting the causal chain anywhere and calling that the cause is arbitrary. Napoleon’s order to cross the Niemen in June 1812 was caused by a thousand antecedent conditions that Napoleon did not control. He felt like he was choosing; but the choice was determined by everything that came before.⁶⁵

5.4 Counterfactual History

The question of whether individuals or structures drive history can be tested (partially) through counterfactual history: what would have happened if a key individual had acted differently, or not been there?⁶⁶

If Hitler had been killed in 1937 — before the Anschluss, before Munich, before the invasion of Poland — would the Second World War still have happened? Structural historians tend to say yes: the conditions for European war (German revanchism, the economic effects of the Versailles settlement, the instability of the Weimar system, the weakness of collective security arrangements) were in place and would have produced some kind of major conflict even without Hitler. Individualists say no: the specific timing, scale, and ideology of the war was Hitler’s; without him, the history looks radically different, perhaps no large European war at all.⁶⁷

There is no way to run the experiment. But the arguments forced by counterfactual questions are valuable precisely because they expose the assumptions embedded in our normal causal narratives.

5.5 The Hierarchy of Causes

Behind the agency-versus-structure dispute sits a more technical problem the lesson has so far left implicit: how does a historian sort the many things that contributed to an event into an ordered explanation? E. H. Carr, in the load-bearing fourth chapter of What is History?, calls this the historian’s professional task — to “establish some hierarchy of causes which would fix their relation to one another,” and “to decide which cause, or which category of causes, should be regarded ‘in the last resort’ or ‘in the final analysis’… as the ultimate cause, the cause of all causes.”⁶⁸ A list of twelve causes of the Russian Revolution earns a second; ranking them earns a first. Crucially, the ranking is interpretive: Gibbon ranks barbarism and religion at the top of Rome’s fall; the Whigs rank constitutional liberty at the top of British prosperity; Marxists rank economic structure. Each ranking is a thesis about what kind of cause matters most, not a discovery written into the evidence.

Richard Evans, responding to Carr in the fifth chapter of In Defence of History, sharpens the conceptual machinery. A cause may be necessary (“if A had not happened, then B could not have happened”) or sufficient (“A happening was enough to make B happen”); historical events are typically overdetermined, meaning “they may have several sufficient as well as necessary causes, any one of which might have been enough to trigger the event on its own.”⁶⁹ So the Sarajevo question divides cleanly. Was the assassination a necessary cause of the First World War? Almost certainly not — the alliance system and mobilisation timetables made some major continental war highly likely, perhaps inevitable, by 1914. Was it a sufficient cause? On its own, no — it would have produced no war without the Austro-Hungarian decision to use it as a pretext, the German blank cheque, the Russian mobilisation, the British failure to deter. Was the war then overdetermined? Carr would say yes — broader trends were always more important than the contingency of one Bosnian Serb’s marksmanship. But Evans warns against letting overdetermination collapse into fatalism: that an event was probable does not mean any particular path to it was inevitable, and the historian’s hierarchy is a defeasible interpretation, not a law.

5.6 Questions to Argue About

• Is the Great Man theory simply false, or does it capture something real about historical causation that structural accounts miss?
• If individuals are shaped by the historical conditions that produced them, in what sense are they morally responsible for their actions?
• Tolstoy’s argument that Napoleon was “history’s slave” seems to imply determinism. If history is fully determined by structural forces, is there anything for an individual to do?
• What can counterfactual history tell us? Is it legitimate as a historical method, or does it confuse empirical history with speculative fiction?

6 Who gets to tell history?

“History is written by the victors.” The phrase is attributed to Churchill, then denied, then re-attributed — appropriately enough for a saying about the unreliability of historical record.⁷⁰ But whoever said it, the insight runs deep. History has traditionally been told by the literate about the literate, by the powerful about the powerful, and by the victorious about their victories. The question of who narrates is not separate from the question of what is known — because the narrators decide what counts as worth knowing, which events count as events, and whose lives count as lives worth recording.

6.1 Wounded Knee: “Battle” or “Massacre”?

The 1890 Wounded Knee event in South Dakota is the canonical Anglophone example of the same naming question the Mohrenkopf raises in Swiss confectionery. On 29 December 1890, United States Army troops of the 7th Cavalry Regiment surrounded a camp of Lakota Sioux at Wounded Knee Creek and, after the discharge of a single rifle in a botched disarmament, opened fire — including with four Hotchkiss rapid-fire mountain guns positioned on the rise above the camp — killing between 250 and 300 Lakota men, women, and children, and pursuing fleeing non-combatants along the creek for several kilometres.⁷⁴ Twenty soldiers received the Medal of Honor for their actions that day. For decades, the event was described in American history textbooks and military records as the “Battle of Wounded Knee.” Dee Brown’s Bury My Heart at Wounded Knee (1970), written entirely from Native American sources and perspectives, was among the first works to use the word “massacre” consistently and to show the event from the perspective of the Lakota.⁷⁵ The question of which word applies is not merely semantic: “battle” implies two armed parties in military contest; “massacre” implies the killing of the largely unarmed. The underlying facts are not in serious dispute among modern historians. A band of Miniconjou Lakota under Spotted Elk had surrendered to the 7th Cavalry the previous evening; on the morning of 29 December the troopers were confiscating weapons. According to the standard reconstruction, a deaf Lakota man called Black Coyote refused to give up his Winchester rifle, which discharged in the scuffle as soldiers tried to take it. The opening shot triggered fire from the surrounding troopers and from the Hotchkiss guns; the shells fell into a tipi camp full of women and children. Soldiers then pursued fleeing non-combatants for several kilometres along Wounded Knee Creek. The Medals of Honor have never been formally revoked, though a congressional resolution calling for their rescission has been introduced multiple times. The linguistic choice between “battle” and “massacre” is a judgement about how to weight that asymmetry — and the same logical structure applies, in the Swiss everyday register, to the Mohrenkopf dispute.

6.2 History from Above and Below

The phrase “history from below” (popularised by E.P. Thompson and other British Marxist historians in the 1960s and 70s) describes the project of recovering the experiences of ordinary people — workers, peasants, enslaved people, women, colonised populations — who have been largely absent from traditional historiography.⁷⁶ This is not merely a matter of inclusivist politics; it is an epistemological claim. If you want to understand the Industrial Revolution, you cannot understand it only from the perspective of factory owners and government ministers. You need to know what it was like to work a 14-hour day in a textile mill at the age of 12.

6.3 Subaltern Studies and Its Limits

The Subaltern Studies collective, a group of Indian historians convened by Ranajit Guha in the early 1980s, set out to recover the history of South Asian peoples from a perspective not filtered through colonialism.⁷⁷ Their central question, taken up by Gayatri Chakravorty Spivak in her influential essay “Can the Subaltern Speak?” (1988), is bleaker than it sounds: can the historically subordinated speak in their own voice at all, or is every attempt to record their perspective already conditioned by the frameworks of the dominant?⁷⁸

Spivak’s answer is uncomfortable: in important ways, no. The subaltern’s voice enters the historical record through documents produced by colonial administration, by mission records, by legal proceedings — all of which impose categories, translate into foreign languages, and select what is worth recording. The attempt to recover “authentic” subaltern voices must contend with the fact that the archive was built by those they oppose.

Trouillot pushes the difficulty one step further. Examining the Haitian Revolution of 1791–1804 — the only successful slave revolt in modern history, which produced the second independent state in the Americas — he argues that the events “entered history with the peculiar characteristic of being unthinkable even as it happened.”⁸¹ His claim is not that no European could perceive the revolt: the August 1791 uprising in the Plaine du Nord was reported in the French press within weeks; the Société des Amis des Noirs (Brissot, Grégoire, Condorcet) had already argued for the political capacity of free people of colour; the civil commissioner Léger-Félicité Sonthonax issued a general emancipation decree in Saint-Domingue on 29 August 1793; and Wordsworth’s 1802 sonnet to Toussaint Louverture (“Thou hast left behind / Powers that will work for thee…”) is exactly the kind of acknowledgment a strong “unthinkability” thesis would predict to be impossible.⁸² Trouillot’s narrower and more defensible claim is that the dominant historiographical and political vocabularies of the era could not accommodate sustained Black political self-government as a normal political event of the kind to be analysed alongside, say, the American or French revolutions; it could be registered only as slave disorder, atrocity, or footnote. The silence in the historical record on this reading is partly archival, partly conceptual: certain things were written down only in the registers (deviation, anomaly, racialised violence) that the dominant categories would admit — and after 1804 the new Haitian state was met with a near-total commercial and diplomatic embargo that did much of the rest of the silencing.⁸³

6.4 The Australian Dictionary of Biography

Since the 1990s, the Australian Dictionary of Biography has undertaken the systematic recovery of Aboriginal Australian histories — an effort that requires developing new methodological approaches, because the evidence does not conform to the conventions of European archival history.⁸⁴ Oral traditions, kinship structures, and material culture become primary sources. The epistemological challenges are significant: how do you cite oral testimony? How do you adjudicate between conflicting oral traditions? How do you recover history from peoples whose archives are landscapes and stories rather than documents?

The challenge is not specific to Australia. It applies wherever colonial history has displaced or destroyed indigenous documentary traditions — which is most of the world.

6.5 Questions to Argue About

• “History from below” is sometimes described as a political project masquerading as a historical one. Is it? And if so, does that compromise its epistemological validity?
• Spivak asks whether the subaltern can speak. What would it mean to answer “no” to this question? Does the answer have practical implications for how history should be written?
• When there is a dispute about whether to call an event a “battle” or a “massacre,” is this a question about facts, language, or politics? Can these be disentangled?
• The Annales school argued that geography and climate are more important historical forces than events and individuals. Does this view risk making human agency — including the decisions that led to atrocities — seem less significant than it is?

7 Can history be a science?

The ambition to make history scientific is old — and old ambitions tend to harden into either embarrassing failures or quiet transformations. The French positivist Auguste Comte, in the 19th century, envisioned a science of society that would include historical laws analogous to physical laws: history as physics, with wars and revolutions playing the role of gravitational collapse and orbital mechanics.⁸⁵ The project has not succeeded in any form that Comte would recognise. But it has not been abandoned, and the most recent attempt is considerably more sophisticated.

7.1 The Annales School’s Ambition

The Annales school — named for the journal Annales d’histoire économique et sociale founded by Lucien Febvre and Marc Bloch in 1929 — represented the most sustained attempt to make history scientific by importing the methods of the social sciences.⁹⁰ Quantitative data, long-term structural analysis, interdisciplinary borrowing from geography, economics, and demography: these were the tools of a historical science that would replace the narrative of events with the analysis of structures.

The achievement is genuine. Emmanuel Le Roy Ladurie’s Montaillou (1975) reconstructed the daily life of a 14th-century Cathar village from Inquisition records with a density of social and cultural detail that narrative history could not have achieved.⁹¹ Braudel’s Mediterranean is a monument of structural ambition. But the programme also has limits: structures cannot account for everything. The particular decisions of individuals in particular moments — the decisions that opened and closed possibilities — do not dissolve into structural analysis.

7.2 Cliodynamics

Peter Turchin, a Russian-American ecologist turned historian, has proposed a discipline he calls cliodynamics — the mathematical modelling of historical dynamics.⁹² His central claim, developed in Ages of Discord (2016) and earlier works, is that political instability in complex societies follows discernible cycles of roughly 50 years, driven by factors like elite overproduction (too many aspirants for too few elite positions) and popular immiseration.⁹³

Turchin offered a prediction in 2010: American society would experience a peak of political instability around 2020. The US Capitol attack of 6 January 2021 is, to his supporters, a confirmation.

The sceptical objections are not trivial. Historical data is fragmentary and imprecisely measured. The categories used (what counts as “political instability”? what counts as an “elite”?) are already interpretive. Selecting which variables to model imports assumptions about which causes matter. And the track record of social science prediction — on recessions, elections, revolutions — is not encouraging.

7.3 What Would Historical Falsifiability Look Like?

Popper argued that the social sciences should be held to the same standard of falsifiability he applied to natural science.⁹⁴ A historical claim is genuinely scientific only if it could, in principle, be shown to be false by evidence. By this criterion, many historical “explanations” are not explanations at all: they are post-hoc rationalisations that could accommodate any outcome.

But applying this criterion rigorously creates problems. Historical events are unique; you cannot repeat the experiment. Historical causes are multiply realised; the same outcome can arise from different configurations of factors. The range of relevant evidence is potentially unlimited. This doesn’t mean history can’t be rigorous — it means that the kind of rigour appropriate to it is different from the rigour appropriate to physics. Collingwood would say: of course. History is not physics. It is the study of human thought, and human thought doesn’t have natural laws.

7.4 Hayden White and the Tropology of Historical Narrative

Hayden White’s Metahistory: The Historical Imagination in Nineteenth-Century Europe (1973) poses the most philosophically radical challenge to the idea that historical narrative mirrors the past.⁹⁵ White argues that historians do not simply select and arrange the facts of the past — they impose narrative form on what is, in itself, a formless chronicle of events. And the form they impose is tropological: it follows the deep structures of literary genres — romance (progress toward an ideal), tragedy (inevitable decline), comedy (reconciliation of conflicting forces), satire (ironic distance from any redemptive narrative).⁹⁶

These are not decorative choices. They are constitutive: the emplotment of the past determines what counts as an event, what counts as a cause, what counts as resolution. Ranke’s narrative of the German nation is Romantic in White’s sense; Marx’s narrative of history is tragicomic (tragedy of class struggle resolving in communist reconciliation); Burkhardt’s narrative of Renaissance Italy is satirical.⁹⁷ The deep narrative structure is chosen before the evidence is examined, not derived from it.

White is not saying historians invent the past. He is saying that the intelligibility of historical narrative — its explanatory coherence — is a literary achievement, not a reflection of how things actually were. The “facts” constrain the story, but they underdetermine it: the same set of events can be emplotted in multiple incompatible but internally coherent ways, all equally faithful to the evidence.

The challenge this poses: if historical narrative is always already emplotted — if the explanatory coherence of a history is a formal achievement rather than a discovery — then what is the criterion for distinguishing good history from bad? White’s own answer: not correspondence to reality (which on his account all competent histories share, to varying degrees) but the quality of the formal choices — their internal consistency, their fit with the evidence, and the historian’s awareness of their own emplotment. White’s critics (Carlo Ginzburg, Perez Zagorin, Allan Megill) reply that this risks letting formal coherence stand in for evidential discipline — that a historian who chooses tragedy over comedy is not making a different kind of choice from a historian who chooses one body of evidence over another, and that treating the choice as constitutive rather than evidential is the wrong philosophical generalisation from the right observation that emplotment exists.

7.5 History as Narrative and the Limits of Prediction

The Malthus and Turchin cases illuminate something beyond the failure of specific models: they are stories about the difficulty of escaping narrative. Malthus told a story about geometric population growth and arithmetic food production; the story had a beginning (population pressure), a middle (insufficient food), and an end (catastrophe).¹⁰⁰ The story was wrong not because the narrative was incoherent but because reality is not narrated — it does not follow the arc of even well-reasoned stories. The Agricultural Revolution, the demographic transition, the Green Revolution were not complications to the Malthus story; they were events outside its logic entirely.

This is the “narrative fallacy” that Daniel Kahneman identifies — the human tendency to construct retrospective stories that make events seem inevitable and predictable, and to extrapolate those stories forward.¹⁰¹ History is particularly vulnerable to the narrative fallacy because it is narrated: we experience it in the form of stories with causes, agents, and outcomes. White’s tropological analysis and Kahneman’s narrative fallacy are arriving at the same point from different directions: the challenge for any discipline that uses narrative is not only to build better models but to recognise that the models themselves are stories, with all the selective, dramatic, causal structure that stories impose — and that this is a permanent condition, not a defect to be cured.

7.6 Questions to Argue About

• If historical patterns like those Turchin identifies are real, does that mean human societies are less free than we think? Or does identifying the pattern give us the means to change it?
• Popper’s falsifiability criterion is designed for the natural sciences. Is it the right criterion for history? What alternative criteria of rigour might be more appropriate?
• The Annales school wanted to replace narrative with structure. But can history do without narrative? Is there something that narrative explanation provides that structural explanation cannot?
• Cliodynamics predicted political instability in the US around 2020. The events of 2020–21 seem to confirm this. Is this genuine scientific prediction, or post-hoc fit? How would you distinguish them?

8 What is history for?

A common professional instinct — articulated forcefully by Geoffrey Elton in The Practice of History and present in the disciplinary self-image of much academic historiography — is to resist this question. History is for knowing. Full stop. On this view, the demand that history serve a purpose — build national pride, teach lessons, provide warnings — is the first step toward distorting it. The instinct catches something real: history put into service of a present cause does tend to be selected and framed by that cause. But it is also incomplete. History is not, in fact, a neutral intellectual exercise pursued in isolation. It is used — by states, by communities, by individuals — for purposes that go beyond the disinterested pursuit of knowledge, and the discipline’s own institutional existence (state-funded universities, school curricula, public commemoration) is bound up with those uses. The question of what history is for is not separable from the question of how it should be done, because the purposes shape the selection, the framing, and the silences whether or not the historian acknowledges them.

8.1 Memory, Identity, and the Nation

Nations are, in Benedict Anderson’s famous phrase, “imagined communities” — and historical memory is one of the primary materials from which that imagination is constructed.¹⁰⁴ The myth of national origin, the founding moments, the defining traumas and triumphs — these are the stories through which a people becomes a people. France’s Revolution; America’s founding documents; Israel’s memory of the Holocaust; the Ottoman/Turkish debates over the Armenian genocide; the Partition of British India in 1947. History is here in the service of identity.

This function is not simply manipulative. Shared historical memory can bind communities, transmit values, and create the solidarity that makes collective action possible. The problem arises when the memory is selective in ways that exclude or demean, when it involves the denial of atrocities, or when it hardens into a mythology that cannot be questioned without appearing to attack the community itself.

8.2 Holocaust Memory and Moral Duty

The Holocaust is the case most directly associated with the argument that historical education is a moral duty — that societies have an obligation to remember and transmit knowledge of historical atrocities to prevent their recurrence. “Never Again” is both a historical claim (this must not happen again) and a pedagogical programme (education about what happened is a means of preventing it).

But the epistemological and ethical relationship between historical memory and future prevention is more fraught than “Never Again” implies. What exactly is the lesson? That genocide is wrong? That is not a lesson that requires historical detail. That modern, educated, European societies can commit genocide? That is a more specific and disturbing lesson. That ordinary people will participate in atrocities when authority structures make it possible and rewarding to do so?

That last connection is most often made in TOK textbooks via Milgram and Zimbardo, but the historical anchor is sturdier. Christopher Browning’s Ordinary Men (1992), drawing on the post-war interrogations of Reserve Police Battalion 101, shows the same pattern in actual perpetrators of mass killing.¹⁰⁶ The battalion was a unit of about 500 middle-aged German order policemen — most too old for army conscription, mostly working-class Hamburgers, not pre-selected Nazi enthusiasts — sent to occupied Poland in June 1942. Their first mass killing came on 13 July 1942 at Józefów, where they were ordered to shoot some 1,500 Jewish women, children, and elderly men at point-blank range. Their commander, Major Wilhelm Trapp, made the unusual offer that any man who did not feel up to the task could step out and be assigned other duties. Twelve out of roughly 500 took the offer; the remaining 80–90 per cent carried out the killings, and the proportion who refused never rose materially over the months that followed.¹⁰⁷ Browning’s central argument is that the killers were neither special sadists nor mere “obeyers”: they were ordinary men whose participation grew through small steps and unit-level peer pressure. Daniel Goldhagen’s Hitler’s Willing Executioners (1996), working largely from the same Battalion 101 archive, drew the opposite lesson — that the killers acted from an eliminationist anti-Semitism specific to German political culture and did not need the situational pressures Browning emphasises.¹⁰⁸ The Browning–Goldhagen exchange is itself a piece of evidence about how the same archive yields different histories — and a useful corrective to the textbook habit of treating Milgram’s lab studies and Zimbardo’s prison experiment as the canonical evidence here. Milgram’s original studies survive scrutiny, though the popular “65% will administer lethal shocks” gloss flattens a wide range of obedience rates across his eighteen variations; Zimbardo’s Stanford Prison Experiment has been substantially impeached by Thibault Le Texier’s 2018 archival reconstruction.¹⁰⁹¹¹⁰

8.3 Truth and Reconciliation

The South African Truth and Reconciliation Commission (1995–2002) and subsequent commissions in Rwanda, Sierra Leone, and elsewhere rest on a specific epistemological claim: that establishing a public, shared record of what happened — truth — is itself a form of justice and a precondition for social healing.¹¹¹ Perpetrators were invited to testify in exchange for amnesty; victims were given a formal space to be heard. The commission produced historical knowledge and used it as an instrument of social policy simultaneously.

The epistemological assumptions here are contestable. Is it really possible to establish a single shared truth about complex atrocities, in which perpetrators, victims, and bystanders all participated? Or does “truth” in this context mean something more modest — a public acknowledgment that certain things happened, even if the full causal account remains contested? And if truth can produce reconciliation, what exactly is the mechanism? Is there evidence that it does?

8.4 The Ethics of Monuments

Monuments are public history — claims about the past installed in shared space. The debates over Confederate statues in the United States, the Rhodes Must Fall movement that began in South Africa in 2015, the removal of Saddam Hussein’s statue in Baghdad in 2003 and the toppling of the slave trader Edward Colston’s statue in Bristol in 2020 — all are contests over whose history is publicly honoured.¹¹²¹¹³

The case for removing monuments is epistemological as well as political: a Confederate statue in a Virginia courthouse square is not a neutral commemoration of the past but an active claim about the present — about whose legacy deserves public honour, about how the Confederacy should be understood. The Southern Poverty Law Center’s Whose Heritage? survey found that the great majority of US Confederate monuments were not installed in the immediate post-bellum period of mourning but in two later waves: the 1900–1920s, coinciding with the consolidation of Jim Crow, and the 1950s–60s, coinciding with the civil rights backlash.¹¹⁴ The Robert E. Lee statue in Charlottesville is a representative case. Erected in 1924 — sixty years after Lee surrendered, well inside the first wave — it became the focal point of the Unite the Right rally on 11–12 August 2017, when several hundred white-nationalist and neo-Nazi marchers gathered to oppose its proposed removal. On the second day a self-identified neo-Nazi drove his car into a crowd of counter-protesters, killing the activist Heather Heyer and injuring thirty-five others.¹¹⁵ The Lee statue itself was finally removed in July 2021 and melted down in 2023. The chronology — installation during Jim Crow, removal in the wake of an avowedly white-supremacist defence of it — supports treating these monuments as racial-political statements of their installation era at least as much as commemorations of the war.

The case against removal is also partly epistemological: erasing monuments is a form of the historical distortion that Orwell described. Better to contextualise, to add interpretive plaques, to let the monument stand as evidence of what people once chose to celebrate.

Both arguments are serious. The question is what you think public monuments are for.

8.5 Questions to Argue About

• Can historical education reliably prevent the recurrence of atrocities? If not, what is it for?
• Truth and Reconciliation Commissions assume that establishing shared historical truth can produce social healing. Is this assumption justified? What would evidence for or against it look like?
• When a community’s identity is bound up with a particular historical narrative, and that narrative is challenged by historical scholarship, who has the obligation to change — the community or the historians?
• Is there a principled difference between erasing monuments (as South African students demanded of Cecil Rhodes’s statue) and erasing historical records (as Orwell’s Winston Smith does at the Ministry of Truth)?

9 Media

• Hilary Mantel, Wolf Hall (2009) and Bring Up the Bodies (2012) — Historical fiction that foregrounds the constructed nature of historical knowledge; Mantel narrates from inside Cromwell’s perspective with no more access to his thoughts than any historian, but makes the epistemological limitations of this a formal feature of the work.
• Leo Tolstoy, War and Peace (1869), Epilogue (Part Two) — Tolstoy’s philosophical excursus on historical causation; one of the most sustained and powerful arguments against the Great Man theory by a novelist working inside a masterpiece of historical fiction.
• Joshua Oppenheimer, The Act of Killing (2012) — A documentary in which Indonesian perpetrators of the 1965 massacres re-enact their killings in the style of Hollywood genre films. One of the most disturbing and epistemologically complex films ever made about memory, violence, and history.
• Toni Morrison, Beloved (1987) — A novel about the aftermath of American slavery, built on the premise that there are historical experiences so extreme they resist direct representation; a meditation on memory, trauma, and the limits of testimony.
• Akira Kurosawa, Rashomon (1950) — The film that gave its name to a permanent epistemological concept: four accounts of the same event, all irreconcilable, none demonstrably false. A 90-minute argument about the insolubility of historical perspectivalism.
• Ari Folman, Waltz with Bashir (2008) — An animated documentary in which the filmmaker, an Israeli veteran of the 1982 Lebanon war, tries to recover his repressed memories of the Sabra and Shatila massacre. The film raises the question of what historical knowledge can mean when the historian is also the eyewitness whose memory has failed.
• Gabriel García Márquez, One Hundred Years of Solitude (1967) — A novel that enacts, in narrative form, the way collective memory operates: mythologised, telescoped, distorted by repetition, yet carrying something essential about historical experience that straightforward chronicle cannot.

10 Bibliography

11 Notes