I've thought of it like an approximation of where the functional parts of the phone were. Thumb is the transmitter and the pinky is the receiver.

That's how Inspector Gadget answered his phone.

More seriously, people picking up the receiver of a Western Electric Model 500 would grip the receiver tightly; extending your thumb and pinky from this grip gets you the "pretend phone" hand mime. Otherwise it just looks like you're holding your fist to your ear. People grip a cellphone quite differently, generally with their fingerpads or the middles of their fingers.

I don't think it is replicating have position, it is just the convenient way to shape the hand so it (sort of) stretches from mouth to ear.

It is common to cartoons and TV shows so some people will have learned it from there, but I'd guess the mine predates those shows. Maybe it when predates TV being common in people's homes.

Holding ones fist up to their ear, as if holding a rotary phone handset, may come across as overtly aggressive.

Or sexual depending on the angle and tightness of fist.