Yeah, so let's be clear about all this.
First: you all probably know this, but power dissipation in anything with resistance goes as P = I*I*R, where I = current, R = Resistance, and P is the power, in watts. Got a hank of wire? It's got resistance. Got a wire clamped between two pieces of metal with a bolt? Copper->OtherMetal will have resistance at the joints, typically more than a straight piece of wire, so of course power is dissipated there.
Then, there's this concept: Thermal resistance. It kind of looks like Ohm's law, but it isn't. Those of us who muck with this kind of thing use the Greek letter, "Theta". I'm not going to spell it out; let's use H. Then,
Tdevice = H * P + Tambient
So, the temperature rise above ambient is the Power being dissipated in something times the thermal resistance. In the above, H can be roughly fixed. One will get one value of H, say, for something sitting in still air (with convection cooling, say) and a much smaller value when one is pointing a fan at it.
Next note: H can add. Suppose that, to get from the joint where the power is being dissipated, one has to go from, say, the joint to the air in the Wall Connector; then, from the Wall Connector case to the outside world. So there'll be a Hj-case and an Hcase-air, and the total H would be the sum of the two.
Final note on this: H can, kind of, divide. Say one has a big honking copper wire. Get one end of it hot and it's going to thermally conduct. (The characteristic is thermal conductivity and one can look it up.) Of course, now with that wire warm, since it's not in the ambient air but surrounded by insulation, conduit, dry wall, and all that, well, the equations get complicated. In any case, for multiple parallel paths H1, H2, H3 to ambient from some point, we get
Htotal = 1/(1/H1 + 1/H2 + 1/H3 ...). Just like EE201 and resistors in parallel.
So, going back to a Wall Connector: Well, there's not much air circulation in there, so whatever the heck H is for the joints in that box to the ambient, it's probably roughly constant.
Now, what is in that box is a temperature sensor. So, roughly, we got Tdevice, or something which is pretty closely related to that being reported to the car/Wall Connector electronics. Here's the problem: Above some temperature in that box or around that joint, the insulation and/or materials in that box are going to deteriorate. This isn't rocket science, everything goes bad if it gets hot enough, including rocket engines. (I've heard the term, "engine-rich exhaust" on some notable.. well, the engine didn't work so well events.) So, reducing the current seems like a right good idea in such a case.
And that makes really good sense. Remember where I said that P = I*I*R? That's a squared term. If 48A is too much, dropping it to 24A is dropping the current by half and the power dissipation by 1/4. And that means that Tdevice - Tambient also goes down by a factor of four, thus saving the day.
So, at this point nothing should be a surprise. Got a loose joint? That'll have higher resistance. Crank up the torque and get it back to spec: If the resistance goes down by half, so does the temperature rise.
But, there's more! On another thread in this forum there was a fellow with a NEMA14-50 (or something like that) and a Mobile Connector. The charging current would drop off. You can probably see where this going: The Tesla Mobile Connector does, indeed, have a thermal sensor in the adapter plug. The socket was all fastened to its wires just fine but it was warm in that garage, so the expected, reduced current to Save The Day kicked in. I suppose that getting a better wall socket from a higher-quality manufacturer might have improved matters somewhat, but this guy's solution was decidedly low tech and effective: He aimed a small fan at the wall socket. Remember what I said about H being variable in the presence of moving air? Worked like a champ.
A hundred linear feet per minute of moving room temperature air can do wonders, vs. the 10 linear feet per minute that one can get with convection. Give it a try? After making sure things are torqued down properly: Safety first!