Outside Edinburgh stands Rosslyn Chapel. Tourists go there now because of Dan Brown's novels and the mystical theories that have accreted around the place. The chapel is incomplete. It's incredibly lavish — every surface covered in carving, dense and excessive. The ground plan is, by contrast, "absurdly simple." It was built in the mid-15th century by William Sinclair, an Earl with enough money to do whatever he wanted.1
The theories about Rosslyn Chapel are extensive. Templars. Holy Grail. Masonic codes. Mystical geometry. Each theory builds on top of the last, and the building has become a kind of palimpsest of interpretive ambition. MacCulloch listens to the theories, considers them, and offers a different reading: "The Earl who built it, the Sinclair Earl, had lots of money, but I think he was also mad. He had enough money to indulge his lunacy to build this really quite silly building, which is beautiful, but isn't is not very sensible."1
He stops to make sure you hear the move he's making. "I've not heard enough people say, 'Look, he's clearly mad.'" The theorizing is taking the place of a simpler explanation: the man who built the building was insane in the terms of his own period, and he had enough money to get away with it.1
That's the methodological move worth filing. Lunacy is a legitimate historical explanatory variable, and historians regularly underuse it because the alternatives — sophisticated motives, hidden meanings, esoteric programs — feel more intellectually serious. They aren't. Sometimes the actor was just mad. Sometimes the historian's job is to say so.
MacCulloch's blunt formulation is worth quoting: "We underestimate the role of lunacy in history. Some people in the past are just mad."1 He gives Rosslyn as one case and Henry VIII as another, but the principle is general. Historical actors include the mad. Mad actors do mad things. The mad things show up in the documentary record. If the historian's interpretive apparatus has no slot for madness, the documentary record looks like a series of inexplicable choices that need to be rationalized into more legible motives.
The rationalization is the failure mode. The Sinclair Earl's overbuilt chapel becomes evidence of Templar conspiracy because the alternative — he was mentally ill and rich — feels too small a story. Henry VIII's last-year personality reversals become the political maneuvering of a calculating tyrant because the alternative — he was in excruciating physical pain and his mood was governed by how he felt that morning — feels too mechanical. The "interesting" interpretation wins by default over the more accurate one.
MacCulloch's correction: include madness in the toolbox. Use it where it fits. Not as a lazy default, but as one explanation among several, available when the evidence supports it and the alternative interpretations are doing too much work.
There's a sub-mechanism in MacCulloch's account that is more interesting than the surface principle. When you bring lunacy into the explanatory apparatus, something specific happens to the historian's relationship with the material. MacCulloch lays out the chain:
"When you see reality clearly, the two words that come to mind are bizarre and funny. And this is what comedians totally get is they look at reality, they're like, that is so weird. And once you get bizarre, you get into mystery, you get into awe, you get into wonder, you get into humility."1
Read the causal chain carefully:
Bizarre → mystery. Accepting that the past contains genuinely bizarre material — buildings built by madmen, decisions made under pain, projects driven by inexplicable obsessions — puts you in a posture where you don't expect the past to make tidy sense. Mystery becomes available.
Mystery → awe. Once you accept that the material is genuinely strange, the strangeness stops being a problem to be explained away and becomes a feature worth contemplating. Awe is the response to material that resists easy capture.
Awe → wonder. Awe gives way to wonder when you start asking what the strange material reveals about the human condition. The mad earl who built Rosslyn isn't just a curiosity; he's a piece of evidence about what humans can do when constraints are removed.
Wonder → humility. The final move: you recognize that your interpretive apparatus is not adequate to the material. The material exceeds the categories. You write your explanations more provisionally, with more space for the irreducibility of what actually happened.
The whole chain is doing methodological work. The historian who admits lunacy as a variable is the historian who can travel the chain from bizarre to humility. The historian who can't admit lunacy keeps trying to rationalize the strange material into something her categories can hold, and never gets past the first link.
Notice that the chain ends in humility, which is what MacCulloch wants the discipline to install. The "may have" syntax he runs at the sentence level is, structurally, the humility the chain produces. Lunacy-recognition is one of the inputs that keeps the humility honest. A historian operating with only rational-actor categories writes confidently because the world her categories describe is fully rationalizable. A historian who has accepted that the past contains genuine madness writes more provisionally because she knows the categories don't cover the territory.
MacCulloch makes a careful distinction that protects the principle from being lazy. Mad and irrational are not synonyms. He points to Henry VIII as an example of irrational-via-illness rather than mad-per-se.
"One thinks of people who are not mad but irrational. And that may be because they're ill. And the obvious example is Henry VIII, who in his later years was constantly in pain. He was also a terrible narcissist. And that is a factor throughout his career, but in the last decade, there is real physical pain."1
A doctor pointed this out to MacCulloch as he was trying to explain "the bewildering turn of events in his last 12 months of life." The doctor said: "Well, remember, he's in terrible pain. You need to understand this complete change of attitude to a particular person by how he feels in the morning. And that really needs to be part of the picture."1
The doctor is doing cross-discipline nerving (see the companion page on that pattern). The point for this page is the distinction: the Sinclair Earl was mad; Henry VIII was ill and in pain. Both cases are non-rational actors, but for different reasons. The mad actor's choices are governed by something internal to their cognition; the ill actor's choices are governed by their body's day-to-day suffering. Both need to be in the historian's explanatory toolbox, and both need to be distinguished from the rational-actor cases where the choices actually do follow from interests and information.
So the toolbox has at least three categories:
The disciplined historian distinguishes among these as the evidence allows. The undisciplined historian treats all behavior as rational-actor behavior and produces convoluted rationalizations for material that needs a different category.
How does the historian actually use this? The protocol is straightforward:
Check whether the rational-actor interpretation strains. When you find yourself producing increasingly elaborate explanations for a historical actor's choices — when the political-maneuvering interpretation requires the actor to be a chess grandmaster, the strategic-vision interpretation requires the actor to be a tactical genius — the rational-actor frame may be doing too much work. Ask whether one of the other categories fits.
Test the mad-actor slot. Would the actor's contemporaries have called the actor mad? (This is the key disciplinary check — mad in the terms of the actor's own period, not mad by current standards.) If yes, and the evidence supports it, the slot fits. Use it. The Sinclair Earl's contemporaries had categories for this, and used them.
Test the ill-actor slot. Are there documentary traces of physical condition that would have governed the actor's day-to-day functioning? Pain, hunger, addiction, untreated illness, neurological events? If yes, those have to be part of the picture, in the doctor-to-MacCulloch sense. The choices the actor made are then partly choices made by their body.
Cross-discipline-nerve the diagnosis. Don't try to do the medical or psychiatric analysis yourself. Consult a doctor, a psychiatrist, a historian of medicine. MacCulloch's doctor friend told him about Henry VIII's pain; MacCulloch's archaeologist friends told him about Bede's underweighted women. The mad-actor and ill-actor slots are best populated through cross-discipline conversation, because the relevant expertise is not in the historian's training.
Write the conclusion with appropriate calibration. "Henry VIII, in his last twelve months, was operating under severe physical pain that documentary sources suggest substantially affected his moods and decisions on a day-to-day basis. This needs to be part of how we read his choices in this period." Not "Henry was crazy." The calibration matters. Lunacy-recognition as discipline is different from amateur diagnosis.
The Rosslyn and Henry VIII cases are compelling, and the principle clearly applies to a range of historical figures that current scholarship tends to over-rationalize. The Roman emperors are an obvious group. So are some of the more florid medieval kings. So are certain modern political figures whose later decisions become illegible if you assume rational-actor logic throughout. The methodological correction is real.
Tension: the principle can be misused to dismiss historical figures who simply had different value-systems than the historian's. "He was crazy" is sometimes a polite way of saying "I don't understand his motivations." The discipline requires the historian to first attempt to understand the actor's value-system on its own terms (the sympathy beat from skepticism-then-sympathy), and only invoke lunacy when the actor's own contemporaries would have used the category. The principle is not license for present-tense psychological diagnosis of the past.
A second tension: the line between "mad" and "ill" is itself a historical and cultural variable. Some periods have categorized as illness what other periods have categorized as madness, and vice versa. The Sinclair Earl's contemporaries had specific theological and humoral categories for what would now be diagnosed differently. The historian using the mad-actor slot has to use the category-system the actor's period recognized, which requires expertise in the period's own conceptions of mental difference.
A third tension lives in the relationship between lunacy as variable and structural explanation. A historian committed to materialist or structuralist explanation may resist the lunacy frame because it locates causation in individual psychology rather than in larger forces. MacCulloch's frame is methodologically pluralist — he wants the toolbox to include lunacy as well as structural variables, with the historian using each where the evidence supports it. The pluralism is itself a methodological position.
Open questions:
Read Le Bon and the territory expands sideways instead of in. He's not interested in the mad individual; he's interested in what happens when individuals aggregate into a crowd. Sensible people become a different creature in the mass. Reasoning drops a level. Vivid imagery overrides argument. The crowd acquires a kind of impersonal madness that the same people, individually, would never enact. MacCulloch is asking the parallel question one scale smaller: what about the single mad actor, the Sinclair Earl alone with his money, the Henry VIII alone with his pain? Both authors are saying the historian's toolbox has to include non-rational explanatory categories. They're working at different scales of the same expansion. The historian who only reads Le Bon will catch crowd dynamics and miss the mad individual; the historian who only reads MacCulloch will catch individual lunacy and miss what aggregates into different behavior at scale.
Now open Foucault on madness and the ground shifts. He's not telling you madness exists or doesn't; he's telling you the category of madness is itself constructed differently across periods. What sixteenth-century Catholic Britain called madness was structured by humoral theory and theological assumption; what nineteenth-century French alienists called madness was structured by clinical psychiatry; what we call madness today is structured by DSM categories that didn't exist a hundred years ago. Apply this read to MacCulloch and you see the vulnerability. Calling the Sinclair Earl "mad" without specifying whose category of madness you're invoking is making an underspecified claim. MacCulloch partly answers — mad in the terms of his own period — but the qualification carries weight only if you actually know that period's category-system. You can't run lunacy-as-variable on autopilot. You have to know what mad meant where and when you're working.
Then read Greene on narcissism and a different category appears next to madness. The Sinclair Earl building his impossibly ornate chapel might not be mad at all — he might be running a grandiose self-narrative that's perfectly coherent from inside his narcissism, even though it looks insane from outside. Henry VIII through Greene's lens isn't mad; he's a narcissistic king whose patterns become extreme under physical pain. The diagnostic difference matters. Madness has a different texture than narcissism, generates different evidence, points to different antecedents. Lunacy and narcissism are neighboring slots in the toolbox; the historian who collapses them into "irrationality" loses the analytic precision both categories were built to support.
Bring Buss into the room and a third slot opens. People often do things that look mad by current rational-choice criteria but are calibrated for environments humans no longer inhabit. The evolutionary mismatch — rational by ancestral design, broken by modern conditions — is neither madness nor narcissism nor illness. It's the cognitive program operating correctly in the wrong world. Henry VIII's coalition strategies might have been evolutionarily sensible in a much smaller political setting and ruinous at the scale of an English kingdom. MacCulloch's three-category toolbox (rational / mad / ill) is incomplete. Buss adds a fourth (mismatched). The richer toolbox catches more of what historical actors were actually doing.
Pull on these threads — the psychiatry literature on historical figures, the philosophy of madness, and the strategic literature on irrationality as competitive advantage.
Psychology: Identity Architecture & Defense Hub — the depth-psychology literature on personality structure, the conditions under which identity organizes around grandiose or paranoid patterns, the somatic factors that interact with psychological structure. The handshake produces a methodological collaboration: the historian who invokes lunacy as a variable benefits from access to the psychiatric literature on what madness has looked like across periods, what its diagnostic features are, how it interacts with social position and resources. The collaboration is mostly absent in current historiography, and the cost is the over-rationalization MacCulloch is correcting.
Behavioral Mechanics: Narcissism Spectrum — the Greene / Lieberman literature on narcissistic patterns. The Rosslyn Chapel case might be narcissism rather than madness; the diagnostic distinction matters for the interpretive consequences. The handshake produces a sharpening: lunacy and narcissism are neighboring slots in the toolbox, and the historian distinguishes between them based on whether the actor's grandiose behavior is intelligible-from-inside-the-narcissism (narcissism) or unintelligible-even-from-inside-the-actor's-perspective (madness). The discipline is to make the diagnostic distinction rather than collapse the categories.
Psychology: Evolutionary Psychology Hub — the EP frame supplies an additional slot: rational-by-ancestral-design-and-mismatched-to-current-environment. Some historical behavior that looks mad is actually mismatch-driven. The handshake produces a richer toolbox: rational-in-context, mad-by-period-standards, ill-and-in-pain, mismatched-by-evolution. The historian working with all four categories has substantially more interpretive precision than the historian using only the first.
Cross-Domain: Anti-Hero's Journey (Solar Consciousness, if exists) or analogous page — the structural pattern of figures whose disturbance produces unusual achievement. The lunacy-equipped historian has the apparatus to distinguish figures whose disturbance was instrumental to their achievements (productive disturbance) from those whose disturbance ended the achievements (destructive disturbance). The handshake produces a typology: not all madness produces nothing; some madness produces lasting structures (Rosslyn Chapel still stands), and the historian's job is to read the relationship between the disturbance and the output.
Psychology: Somatic-Trauma Theory Hub — the somatic literature provides the body-state variant of the lunacy slot. Trauma-driven behavior is not madness in the period-recognized sense, and it isn't physical illness in Henry VIII's sense either. It is a third category: choices governed by autonomic nervous-system state, by triggered defensive patterns, by dissociation, by hypervigilance. Historical figures with serious trauma histories — and many of them had such histories — would have made decisions partly dictated by the body's threat-response system rather than by deliberate calculation. The handshake produces a fourth slot in the historian's toolbox alongside rational-actor / mad-actor / ill-actor / mismatched-actor: trauma-driven actor. The four-slot toolbox captures more of the actual historical record than the standard rational-actor frame, and a historian working with all four categories is operating with substantially more interpretive precision.
Behavioral Mechanics: Manipulation & Influence Hub — manipulators often deliberately destabilize their targets' rational-actor functioning, producing behavior that subsequent historians read as inexplicable. Sleep deprivation, isolation, calculated humiliation, fear-induction — these techniques shift the target into states where rational-actor analysis fails. A historian who knows the manipulation literature can identify when a historical figure's behavior makes sense as the predictable output of someone in a manipulation-induced state. The handshake produces a methodological refinement: lunacy-as-variable should include lunacy-as-induced, where the actor's choices are governed by a state another actor has deliberately created. This applies to interrogation contexts, abusive relationships, and certain political pressures throughout history.
The Sharpest Implication
The principle, generalized, says that the rational-actor model is the historian's most common analytical failure mode. The temptation is constant: every actor is treated as if their choices follow from interests + information, modulated by ordinary cognitive limitations. This frame is comfortable. It's also wrong about a substantial fraction of the historical record. Real history includes madmen, sick people, people governed by their bodies' pain or hunger or addictions, people whose values are not your values, people whose situations you cannot reconstruct. If MacCulloch is right, the disciplined historian writes with much smaller confidence than the standard rational-actor frame supports — because the frame's confidence comes partly from how much of the actual material it filters out as inexplicable. The third-wire reading is that the same correction applies to your reading of present-tense people: the colleague whose behavior makes no sense, the family member whose decisions seem self-defeating, the public figure whose choices appear inexplicable. Some of them are mad. Some of them are ill. Some of them are mismatched. Not everyone you're trying to understand is a rational actor, and the assumption that they are produces increasingly elaborate misreadings of behavior that has a simpler explanation.
Generative Questions