I asked my kids while we were watching Harry Potter and the Order of the Phoenix: who’s the most evil villain you can think of? Riley said Pennywise.
Ugh.
Here’s the thing about Pennywise, or the Joker, or any villain whose evil exists purely for its own sake: they’re boring. Not as horror – they’re effective enough at making you check the shower curtain. But as moral objects they’re closed loops. Ask “why?” and you get nothing back. It’s like asking why a fire burns. It just does. Good for jump scares; lousy for moral philosophy. Evil with no interior is just weather with teeth.
But the question stuck with me, because when I turned it on myself, my gut instinct felt just as wrong – just wrong in a more interesting way. My first answer was something small. Not a Voldemort, but a bureaucrat. A petty functionary. The kind of villain who smiles while she hurts you. Umbridge. And I’ve noticed I’m not alone in this; ask the question broadly and people reliably skip past the apocalyptic and land on the familiar.
I think that instinct reveals something uncomfortable – not about the villains, but about us.
We don’t actually rank villains by the harm they do. We rank them by how much they make our skin crawl, and our skin crawls most at what we recognize. Nobody’s met a Voldemort, but everybody’s had a boss who lords over their employees like a personal fiefdom and makes them file TPS reports on the weekends. The political philosopher Hannah Arendt called this the “banality of evil” – the idea that atrocities aren’t always driven by monsters, but administered by unremarkable people following procedures. The petty villain is this concept with a face, and it hits harder than mythological evil ever can.
But hateability and moral weight are not the same thing. I’ve been turning over a taxonomy in my head that I think pulls them apart.
A Taxonomy of Evil
Here’s what I’ve landed on: several classes of evil, ordered not by how much they make your skin crawl, but by how much damage they can actually do.
Evil as Petty Authority
The archetype: Dolores Umbridge (Harry Potter), Ambrose Jakis (The Kingkiller Chronicle)
This is cruelty exercised within existing structures. The petty authoritarian doesn’t build anything; they inhabit systems that already exist and find the spaces where they can hurt people without consequence. Umbridge needed the Ministry. Jakis needed his family’s wealth and the University’s political hierarchy. Remove the structure and they’re just unpleasant people at a dinner party.
And honestly? Jakis is kind of pathetic. Kvothe makes a mockery of him. The song, the public humiliations – Jakis can’t win on his own merits and everybody knows it. That’s the tell. Petty authority villains are tools in both senses of the word. They’re effective only within the machine that houses them.
This evil is horrible to experience, but it’s bounded. It requires permission. It operates within the rules, just the worst possible interpretation of them. It’s the lowest tier not because it doesn’t cause real suffering – it does – but because it can’t scale beyond the reach of the institution that hosts it.
Evil as Appetite
The archetype: Croup and Vandemar (Neverwhere), Mr. Dark (Something Wicked This Way Comes)
This is predation without ideology. Neil Gaiman writes Croup and Vandemar as something ontologically wrong; ancient, gleeful, and utterly without interiority. Croup is verbose and mannered; Vandemar is blunt and hungry. Together they function like a single predator split into two complementary lobes: the one that talks to you and the one that eats you.
Mr. Dark operates similarly but through seduction rather than violence. He doesn’t take; he offers, and lets the offer destroy you.
These villains are terrifying but not scalable. They operate outside institutions – they can appear at will, tempt, destroy – but they’re still contained by their own nature. They can’t recruit. They can’t build movements. Appetite is solitary by definition; it consumes, but it doesn’t organize. A knife only cuts what it touches, even if the knife can walk.
Evil as Craft
The archetype: Melisande Shahrizai (Kushiel’s Dart)
Intelligence weaponized for personal ends. Melisande is the most beautiful woman in a nation that worships beauty, and the most politically brilliant mind in a court full of schemers. She betrays her own country, orchestrates an invasion, and manipulates the protagonist into being an unwitting pawn – all while being so charismatic that Phèdre, the woman she’s actively destroying, can never entirely stop being drawn to her. What makes this class distinct is that the destruction itself is the art. The villain takes pleasure in the elegance of the scheme, not merely its outcome.
More dangerous than appetite because ambition pushes the edge of malevolence further than hunger ever could. Melisande doesn’t just want to consume; she wants to reshape. She topples governments. But it’s still individual in scope; it’s one brilliant mind working one intricate plan, and when that mind is removed, the evil stops.
Fernand Mondego from The Count of Monte Cristo sits at the boundary between this class and the next. He schemes like Melisande – the betrayal of Dantès is pure craft – but then he builds on it, parlaying his treachery into lasting social position, military rank, and political power. He’s a craftsman who succeeds well enough to become an institution. Characters like this are worth noting because the taxonomy isn’t rigid; the most interesting villains often bridge categories.
Evil as Competence
The archetype: Tywin Lannister (A Song of Ice and Fire), Sadeas (The Stormlight Archive)
Amoral pragmatism. Tywin doesn’t enjoy cruelty; he just doesn’t let it slow him down. The Red Wedding isn’t sadism; it’s project management. Sadeas uses Dalinar’s bridge crews as cannon fodder, and when he betrays an ally on the battlefield, it’s not personal – it’s strategic calculation. Neither of them would describe themselves as evil. They’d describe themselves as realistic.
This is more dangerous than craft because it builds dynasties, not just schemes. Tywin’s evil doesn’t end when he dies; it reverberates through the institutions and precedents he created. He’s also the most recognizable villain on this list if you’ve ever worked in a large organization. The executive who restructures three departments to eliminate one problem employee and calls it efficiency.
Evil as Ideology
The archetype: Judge Claude Frollo (The Hunchback of Notre Dame)
Self-justifying, systemic, recruitable. Frollo’s résumé is staggering: attempted infanticide, ethnic cleansing of the Romani, religious persecution used to justify mass murder, and sexual obsession externalized as the victim’s sin. He does all of this while believing he’s righteous.
“Hellfire” might be the most psychologically sophisticated three minutes Disney ever produced; a villain song about a man blaming God and a woman for his own arousal.
This is where evil starts multiplying, because ideology can be taught. Petty authority needs an institution. Appetite needs proximity. Craft needs genius. Competence needs capability. But ideology just needs a willing ear. It’s a contagion that scales with population. Historically, this is how the worst real-world atrocities actually work; not through the genius of one schemer, but through a framework that lets ordinary people justify extraordinary harm.
Evil as System
The archetype: Matthew Sobol (Daemon), The Borant Corporation (Dungeon Crawler Carl), The Ministry in Brazil
Evil that no longer requires a villain to operate.
Sobol is already dead when his evil begins. His code – the Daemon – reshapes society according to his design, and there’s no one to negotiate with, no one to persuade, no one to defeat. Just architecture executing intent. The Borant Corporation takes it further; genocide as a content production pipeline, mass extinction as a ratings play. The executives making these decisions have probably never personally harmed anyone. They don’t need to. The system does it for them, and the system has metrics. Terry Gilliam’s Brazil imagines a world where the system has become so total that the question of who’s running it doesn’t even have an answer. Nobody is. It just runs.
This is the most dangerous class because there’s no head to cut off. Kill Frollo and his ideology might die with him, or it might not. But a system isn’t a body with a brain; it’s a process made of people and procedures, and it keeps operating because operating is what it does. It doesn’t need a villain at the helm. It doesn’t even need to be aware that it’s causing harm. It just needs to keep running.
The Hateability Inversion
Here’s what bothers me about my own gut reaction: the most dangerous evil is the least viscerally hateable.
I despise Umbridge. Croup and Vandemar are deliciously loathable. But Tywin? I catch myself admiring him sometimes. And the Borant Corporation is so abstract it barely registers emotionally; it’s hard to feel rage at a content delivery pipeline.
I instinctively rank evil by how much it makes my skin crawl, and my skin crawls most at what I recognize. Petty authority is familiar, and easy to dismiss. Systemic evil is theoretical. But ideological evil – the Frollos – that’s the one that keeps me up, precisely because it wears the mask of conviction. So I put the ideologues near the top and barely think about the system that empowers them in the first place.
That inversion – the most dangerous evil being the least emotionally accessible – might be the most important thing this taxonomy reveals. It certainly made me rethink my Dumbledorian deus ex machina explanation for my choice.
The Perpendicular Axis
These classes describe what evil does. But as I worked through this, I realized there’s a second axis – perpendicular to the first – that describes how evil relates to itself:
Self-aware evil. Melisande, Tywin. They know what they are. They’ve decided the cost is acceptable or simply don’t care.
Self-deceived evil. Frollo. Knows something is wrong, constructs an elaborate justification to avoid facing it. “Hellfire” is about a man almost achieving self-awareness and then flinching away.
Self-righteous evil. Thanos, Frollo again (in a different light). Genuinely believes harm is the moral choice. The most dangerous variant because it’s immune to conscience; conscience is what’s driving it.
Self-righteous evil is chilling precisely because from the inside it looks identical to heroism. The only difference between Thanos and a protagonist making hard choices is whether the author frames the sacrifice as justified. And Thanos’s evil is specifically self-righteous ideology; he has a whole cosmological framework about balance and resources that makes his genocide feel (to him) like an act of mercy. He grieves the cost. That’s what makes him more terrifying than a Pennywise.
This is worth sitting with, because a sufficiently well-crafted villain can be reframed as the hero with surprisingly little effort. The Ruhar in Craig Alanson’s Expeditionary Force invade Earth – an act of unambiguous evil from humanity’s perspective. But their motivation is to make humans an unappealing target for the Kristang, a species that would enslave us permanently. From the Ruhar’s perspective, they’re performing a rescue operation that happens to involve bombing Columbus Day parades. Whether that’s heroism or villainy depends entirely on which end of the telescope you’re looking through. The most dangerous evil isn’t just immune to conscience; it might genuinely be indistinguishable from good, viewed from the right angle.
You can cross-reference the two axes. Frollo is ideology plus self-deception. Tywin is competence plus self-awareness. The Borant Corporation is system plus… what? Indifference? They’re so far removed from the harm they cause that the question of self-awareness barely applies. That’s what makes systemic evil so hard to fight; there’s no conscience to appeal to, deceived or otherwise.
Where Evil Ends
I realized early on that this taxonomy needs a boundary: evil requires intent. Not necessarily active intent, but intent embedded somewhere in the design.
The protomolecule in The Expanse isn’t evil. It has purpose the way a virus has purpose; it executes design, but there’s no mind choosing to harm. It’s catastrophic, but catastrophe isn’t evil. A hurricane isn’t evil. A bear isn’t evil. Evil requires a framework in which harm is meaningful to the one causing it, even if only at the design stage.
This is where it gets genuinely hard to draw lines. The Daemon is just code running. Sobol is dead. Where is the intent living? If you say “well, Sobol originally intended it,” someone can point out that the Builders originally designed the protomolecule. The distance between creator and running system is the same in both cases.
I think the distinction is that the Daemon’s outputs are recognizable as human moral choices, even if no human is currently making them. You can look at what the Daemon does and trace it back to a value system – Sobol’s. The protomolecule’s outputs are alien; you can’t map them to any framework you’d recognize. Evil, in other words, has to be translatable. You have to be able to point at the harm and say “someone chose this, and they understood what they were choosing.” When you can’t do that anymore, you’ve crossed the line from evil into something else – something dangerous, maybe even more dangerous, but not evil. Just the universe being indifferent.
The Mirror
With all the monsters now on the page, and the analysis at arm’s length, I get to feel clever about the taxonomy. It’s a comfortable exercise.
So let me ruin that.
What kind of evil would I be?
I build systems. I design processes that outlive individual decisions. I think in terms of architecture, incentive structures, and creative precision. When I encounter a problem, my instinct isn’t confrontation; it’s restructuring the system so the problem resolves itself. I route around obstacles rather than through them. (Which, incidentally, is why I’m rather bad at direct confrontation – I’m not the petty bureaucrat looking to get kicks from lording power over anyone. I just don’t have the practice.)
If I turned evil, nobody would notice for a while. There wouldn’t be a moment of cruelty to point at. There’d be a system that produced outcomes that happened to benefit me at others’ expense, and it would look like effective and clever planning. The intent would be embedded in the design, and I’d have a rational justification for every piece of it.
I wouldn’t be Frollo. I’d never need to deceive myself. I wouldn’t be Umbridge; petty authority would bore me. I’d be somewhere between Tywin and Sobol – competence that hardens into system, with a framework in which the harm is meaningful but acceptable.
The scariest evil isn’t the kind that’s alien to you. It’s the kind that would use skills you already have, values you already hold, just pointed slightly differently.
I find Umbridge abhorrent, sure – but mostly she’s just not worth the time to correct. The villain that truly gets under my skin is Frollo. The skewed idealist. The one who’s certain he’s right, and has built an entire moral architecture to prove it. Maybe that bothers me so much because it’s the photographic negative of what I do; the same impulse toward systems and frameworks and elegant justifications, just aimed at something monstrous. Tywin’s evil I can at least see clearly for what it is. Frollo’s? That’s the kind that could fool even the person doing it.
Riley said Pennywise. I should probably be grateful he didn’t say “you, Dad, but with slightly different priorities.”
If you’re honest with yourself, you know where you’d fall too.
Comments
To comment on this post, search for this URL in your ActivityPub client (such as Mastodon):
https://asymptomatic.net/posts/2026-03-24-conditions-of-evilNo comments yet. Be the first to reply via ActivityPub!