How a Castle Actually Worked

Most castle explanations start with a cutaway drawing and label the parts. That tells you what things are called, which the glossary does better, and not what they were for.

This page reads a castle the way an attacker would have had to: from the outside in, one obstacle at a time. Done that way, two things become obvious that a labelled drawing hides. The entrance, not the wall, is where all the ingenuity went. And a surprising amount of what a castle did had nothing to do with fighting.

When UNESCO inscribed the castles and town walls of King Edward in north Wales, it listed the vocabulary of the type in a single line: barbicans, drawbridges, fortified gates, chicanes, redoubts, dungeons, towers and curtain walls. It is a good list because of the order. Five of the eight items are about the entrance.

An attacker coming up to a serious castle met them roughly like this. First the ground itself, cleared and sloping away so there was nowhere to shelter. Then the ditch, wet or dry. Then a barbican, an outwork sitting in front of the gate whose whole job was to stop anyone approaching the gate straight on. Then the bridge, which lifted. Then the gate passage, and inside it a portcullis, then murder holes in the ceiling, then frequently a second portcullis, so that a party who rushed the first one found themselves shut in a stone box with the ceiling open above them.

Only then the courtyard, and only then the keep.

Nothing in that sequence had to succeed on its own. Each element bought time and forced the attacker into a place that was already covered from above. That is the whole design philosophy, and it explains why the gate absorbed so much of the money.

Look at the base of almost any serious curtain wall and you will find it thickened and sloped outwards, a feature called a batter or a talus. It solves two problems with one shape.

The first is undermining. The classic way to bring down a wall was to dig beneath it, prop the tunnel with timber, and burn the props. A thick sloping base means far more material to get through and a footing that is harder to cut. The second is that anything dropped from the parapet strikes the slope and bounces outwards, into whoever is standing at the foot of the wall, rather than landing straight down.

The same logic explains round towers. A square tower has corners, and a corner is where a miner attacks, because removing it drops two walls at once. A drum tower has no corner to remove and its curve deflects a strike that would have hit a flat face square on.

The top of the wall is more sophisticated than it looks too. English Heritage describes the arrangement precisely: the battlements allowed soldiers to shoot arrows and crossbow bolts, and hinged wooden shutters between the stone merlons would drop shut to protect them while they reloaded. The gaps we see now were not simply open. In wartime a timber gallery, a hoarding, could also be built out from the wall top, so that defenders could drop things directly down the face without leaning over it.

Film sieges are assaults: ladders, ropes, a breach, a charge. Real ones were mostly arithmetic. An attacking army camped around a castle is also being fed, in hostile country, usually in the campaigning season, and often for months. A siege was a contest between two larders.

Which puts the well, the cistern and the undercroft near the top of the list of things that mattered, and puts the battlements a good way down it. A castle with a deep well inside the walls could sit out an army. A castle whose water came from outside could not, whatever the parapet looked like.

The other consequence is that fire was the real enemy. Castle floors, roofs, hoardings, doors and stores were timber. This is why English Heritage's account of siege weapons ends with the observation that quietly reverses the whole image: it was more common to pour water to put out fires than to dump unpleasant objects on attackers. The defenders were mostly firefighting.

Garrisons were small, too. The picture of hundreds of men on the walls belongs to the cinema; a working castle was held by a few dozen at most, which is why the design leans so heavily on making a small number of people effective in a confined space.

Read only the defensive features and you will misunderstand the buildings, which is roughly what happened to castle studies for most of the twentieth century. Two pieces of evidence from the bodies that protect these sites point the other way.

Historic England's standard description of the roughly six thousand moated sites in England says most of them were prestigious residences, with the moat intended as a status symbol rather than a practical military defence, and that they peak between about 1250 and 1350. And the surviving licences to crenellate, around 463 of them between 1199 and 1600, went overwhelmingly to minor knights for modest manor houses rather than to great lords, which makes them a mark of rank rather than a security control.

Then there is the appearance. Cadw describes Conwy, one of the most formidable castles in Europe, in a way that no photograph prepares you for: this darkly forbidding fortress was originally white, and a thirteenth-century visitor would have seen gleaming white walls, heraldic banners and painted shutters. The lime render kept the water out and it made the building visible from a very long way off. Both jobs at once, again.

So the honest summary is that a castle is a defensive building doing three things simultaneously: keeping people out, keeping a household, and telling everyone within sight who the household belongs to. The parts do not sort neatly into military and domestic, because they were not designed to.

Historic England's monument type thesaurus, which is the classification archaeologists actually use, lists the narrower types under castle. It is a useful list to have in your head, because it is the vocabulary a site interpretation panel will use without explaining.

Two of those entries are worth pausing on. An adulterine castle is one built without permission, a category that only exists because permission was a thing, which loops straight back to the licences. And a siege castle is a castle built to attack a castle, which tells you everything about how long these operations lasted: if it is worth building a fort to conduct your siege from, you are not expecting to be finished by autumn.

Five things turn a walk round a ruin into a reading of it, and none of them needs a guide.

Stand in the gate passage and look up. The grooves in the side walls held the portcullis. If there are two sets of grooves, look at the ceiling between them: those openings are murder holes, and you are standing in the trap.

Find the batter at the base of the wall and work out which way a dropped stone would bounce.

Count the corners on the towers. Square usually means earlier, round usually means the designers had started worrying about mining.

Look for the sockets. A row of square holes high on the outer face once carried the beams of a timber hoarding. Where those holes have been replaced by stone brackets, you are looking at permanent machicolation, which is the same idea built to last.

Find the well. It is usually in the courtyard, it is usually unremarkable, and it is the single feature most likely to have decided whether the place held.