Way back in historical times, in the US, a variety of wiring systems were used, but quite generally the earliest of systems consisted of two wires - a "goesinto" wire and a "goesouta" wire. Wiring in homes and elsewhere was often done with "wire and post" in which bare (or only sometimes insulated) wires were strung post-to-post on glass or ceramic insulators. A little later, it became common to use insulated wire, but the insulation consisted mainly of cotton or linen cloth strips wrapped around the wire and saturated with coaltar pitch or shellac. This is a "not very durable" insulation, but the standoff height of the posts made it relatively unimportant how good the insulation was, as long as it was strung in unoccupied walls and attics etc.

In the era when wire and post installations were done, there was no universal attempt to "ground" anything. Wiring of this kind is still occasionally found still in use by people restoring "historical" homes and other buildings. In areas with "modern" building codes installation of this kind of wiring disappeared prior to 1940 or so. Legal or not, such methods were occasionally used in rural/farmstead homes as late perhaps as the early 1950s ("Uncle E" wasn't really all that up on safety, but had no difficulty getting the components ca. 1948 out on the farm.)

Somewhat later, the use of 2-wire assembled wiring, in which two copper conductors were molded into a single insulating unit became common, and in most areas became required by building codes. At about the time that the molded wire appeared, most codes specified that one side of the circuit must be "grounded" at the "fuse box." In the earliest period when grounded circuits were specified, receptacles where you'd plug in an appliance had two identical parallel slots, so you had no real way of knowing which side was the "hot" slot and which side was the "ground" slot. Systems of this kind were common until sometime perhaps in the late 1940s or early 1950s, and the symmetrical receptacles are still fairly frequently found in "modern" homes (although the mains etc generally will have been brought sort of up to date, - usually - or sometimes - or not).

Recognition that knowing the difference between hot and ground sides of the circuit had some importance led to the use of two-slot receptacles in which the "ground side" slot was a bit wider than the hot side slot. Plugs on appliances where it mattered had a wider blade on one side, so the plug could only be inserted one way. Appliances where it didn't matter had the same size blade on both sides of the plug, and could be inserted either side up. Homes constructed after about 1950 or so, perhaps a bit earlier, will generally have (or had originally) the assymetrical two-hole receptacles, although it's not uncommon to find them "wired backwards."

Sometime in or before the 1950s someone made the remarkable discovery that if the "ground" wire carries a current it doesn't have the same voltage at both ends, and hence at one end or the other it's not really at "ground voltage." Three-hole receptacles became mandatory where there were building codes, although it was probably in the middle to later part of the 1950s before this requirement was generally in place. The wider slot became theoretically the "neutral" connection, and carries the "return" current. The narrow slot is (supposed to be) the "hot" connection. (Miswiring is fairly common.) The third is a round hole, and is supposed to be connected to a true "earth ground." The "ground" wire should never have a measurable current in it except in the case of a malfunction.

The GFI receptacles in common household use in the US generally assume that in order for there to be a "ground line current" there must be a difference between the currents on the "hot" and "neutral" lines, and the difference between these two currents is sensed to determine when the GFI device should trip and interrupt the circuit. By using the difference between two currents, they're insensitive to variations in line voltage, and they can be made very sensitive.

(Sensitivity is good enough that an inductive device like my dog clippers will frequently trip one. The current on both lines is the same, but isn't exactly the same at exactly the same time, during the brief "storage buildup" time in the vibrator coil.)

The "RCD" name may refer to "Return Current Detector" which is a nomenclature now used rarely in the US. What would be called an RCD interruptor would be tripped by the presence of a current in the "third wire." This would respond to an internal short circuit in an appliance, but does NOT reliably trip the circuit if an external return (as through the bath water) is the cause of the fault, where the differential current detection method would. (It's quite possible that other users have updated the method and just retained an archaic terminology.)

Note that dates given here are approximate, based on what I've seen done mainly in one local area. And perhaps not everything that I've seen has been strictly legal. Uniform US National Electrical Codes have been around for a long time, but universal adoption and enforcement is a fairly recent innovation in some places, so there has been a lot of variation across the country. Comments here also apply only to "utility" circuits, sometimes called "lamp circuits." Other rules apply to "heavy appliances."

John