Picture a lamp with its switch off and its bulb unscrewed. You reach in to clear out a broken bit of glass. The lamp is off. The room is calm, and nothing about the moment suggests danger. Whether your fingers are safe depends on one choice, made long before you owned the lamp, about which of two wires the switch was allowed to cut.
That choice is easy to miss. Most people never learn that a switch could have been wired to interrupt either wire, and that nearly every code and product standard insists on one of them.
Two wires, one of them is already "ground"
A standard 120-volt circuit has a hot conductor and a neutral conductor. The utility's transformer produces the voltage between them. The key fact is that the neutral is deliberately tied to earth at the service, at the main panel. That bond holds the neutral near the same potential as the dirt, the plumbing and the concrete floor.
So the neutral isn't "the return wire" in any special sense. It's the wire kept at about zero volts relative to everything you might be touching. The hot wire swings 120 volts away from that reference, 60 times a second.
You get shocked by a voltage difference across your body. Standing on the ground, you are already close to neutral's potential. Touch the hot and the difference is the full 120 volts. Touch the neutral and there is almost none.
That is why the switch belongs on the hot. When it opens, the load side of the circuit is cut off from the 120-volt source. The wire running to the fixture, the socket and the terminals sits at roughly ground potential, and a hand on any of them has almost nothing to push current through it.
What a switched neutral does to a "dead" fixture
Now reverse it. Put the switch in the neutral and leave the hot permanently connected to the fixture.
The lamp still goes dark when you flip the switch, because a circuit needs a complete loop to carry current. With the loop broken anywhere, the filament or LED driver stops conducting, and nothing looks wrong.
But the hot wire is still attached to the fixture, and the fixture is still sitting at 120 volts relative to ground. It is now a live conductor connected to the rest of the world through only the resistance of the lamp itself. If your body offers a path to ground, current can run through the load, through you, and into the earth. A cold incandescent filament can have a resistance of a few tens of ohms. That is almost nothing next to the resistance of a person standing on a damp basement floor.
The design intent behind a standard lamp socket is built around this. In a screw-base socket, the threaded shell is the neutral and the small center contact at the bottom is the switched hot. The spot your fingers naturally touch while unscrewing a bulb is meant to be the low-potential side. Wire it backward and you have placed the energized part exactly where hands go.
The same logic is why plugs are polarized. The wider blade is the neutral, so that the hot arrives on the narrow blade and a well-made appliance can route its switch and fuse onto the hot side and keep the exposed metal shells on the safe one. A lamp wired with the polarity swapped works perfectly, and that is the trouble. Nothing announces the hazard.
What the code actually says
The NEC states this in plain terms. NEC 404.2(B) says that switches or circuit breakers used as switches must not disconnect the grounded conductor of a circuit, with an exception: a switch may do so only if it simultaneously disconnects all conductors of the circuit, or if it is arranged so that the grounded conductor cannot be opened unless all the ungrounded conductors are opened first. That is why a two-pole switch, which cuts both legs together, is fine, and a single-pole switch in the neutral is not.
The code treats overcurrent protection the same way. NEC 240.22 says no overcurrent device may be placed in series with a grounded conductor, with narrow exceptions: one that opens all conductors of the circuit at once, or where the device is required for motor overload protection. The reason follows the same shock physics. A fuse or breaker that opened only the neutral on an overload would clear the fault current path while leaving the hot wire energized all the way to the load. A breaker's whole job is to take the source of voltage away from the wiring downstream. Cutting the wire that is already at ground potential doesn't do that.
There's a second, more subtle reason the neutral isn't fused. If a neutral were opened while loads were still running on the hot, the voltage at the load would float, and on shared-neutral arrangements it could move well away from 120 volts. You can read about that failure on its own terms in the open-neutral symptoms most people never connect to the panel.
Why this still fails in real houses
You might assume this is settled history. It isn't, for a few ordinary reasons.
- Old or improvised wiring. Cheap or older fixtures, DIY lamp repairs and rewired cords can end up with the switch on the neutral side because the cord's two conductors look identical.
- Switch loops. When power arrives at a light box and a cable runs down to a switch, the cable's two conductors can carry the hot out and back. A mistake in identifying them can leave the fixture permanently hot.
- Backwards outlets. A receptacle with reversed polarity places the hot on the wide-slot side, which defeats the polarized plug's protection entirely.
- Assumptions about "off." A switch in the off position tells you nothing about which conductor it opened. The toggle looks the same either way.
None of these will trip a breaker or announce themselves. The system just waits for someone to touch the wrong part at the wrong time.
Your next moves
- Check one lamp today. Unplug a lamp and look at the plug. The wide blade is the neutral. On a two-wire cord, the neutral side has ribbing or a printed marking, while the smooth side is the hot. Then look in the socket: the switched hot should reach the small center tab, and the neutral should reach the threaded shell.
- Test fixtures with the switch off. Before replacing any fixture, turn the wall switch off and use a non-contact voltage tester at the fixture's wires. If either wire still reads live with the switch off, the switch is probably interrupting the wrong conductor or the loop is miswired. Stop there.
- Prove your tester first. Check the tester on a known live source, test the fixture, then check it on the known source again. A tester with a dead battery will tell you what you hoped to hear.
- Mark switch-loop conductors. If you find a white wire doing hot duty in a switch loop, wrap it with black or red tape at both ends so the next person isn't guessing.
- Test the receptacles in a room you use. A $10 receptacle tester shows reversed polarity in a second. If a light comes up wrong, don't use the outlet for anything with an exposed metal shell until it is corrected.
The calculation is easy, the habit is not
The underlying idea is simple enough to carry anywhere: shut off the wire that carries the danger, and verify that it is the one you shut off. Once you start thinking this way, you see that "off" is a claim about one specific conductor, not a feeling about the room.
That is the kind of check Voltly is built to keep close at hand. It's an offline field calculator and NEC reference for the work between the wire and the panel, so the code section behind a rule like 404.2(B) is a few taps away on a jobsite with no signal, next to the voltage drop, box fill and ampacity tools you already use. If it would be useful in your toolbox, you can find it at https://voltly.lumenlabs.works.