If there are any metal parts on the speaker or the baffle plate that are not part of the audio circuit, it is a good idea to ground them to reduce RFI pickup.
I don't want to go on and on about this, but there are two poles on a speaker, let's say + and - . The minus will go to the ground anyway. So, why not short the metal parts that are not on the audio circuit with the - and spare one cable? They will be shorted in the jack anyway. That way the metal parts will be connected to the shield and will be grounded as well just as it is the case in the current setup. Why do we need three instead of two? It doesn't make any sense electrically unless it has a function in the balanced setup. (Don't get me wrong I'm just trying to understand the difference.)
You don't want to go on but you will anyway
In cheap headsets, it is exactly as you describe. One ground (which is also the shield if there is one), one left + and one right +. All metal parts are hopefully grounded.
In better headsets, like the T1, you have ground (the shields, for there are two) , left +, left -, right + and right -.
What, you say, is the benefit of this?
Remember that the same current is flowing in both sides of the circuit. Remember also that all currents have associated electromagnetic fields. These fields do not respect dielectrics to any great degree and therefore
interact with each other.
When both speakers share a common ground wire all the way to the headset, the left and right currents are comingled in the same wire and separation is reduced. Having separate + and - wires for each channel and shielded from each other by a grounded screen increases separation and subjectively improves soundstaging.
The speaker in a headphone converts electrical signals to movement. It also equally well converts movement into electrical signals. The moving parts of a speaker don't start and stop exactly as the signal requires. The better the speaker, the closer the correlation but there is always some back generation as the speaker oscillates to a standstill. This unwanted generation is damped by the output impedance of the amplifier. The lower the output impedance of the amp, the greater the back generated current and the faster the unwanted energy is dissipated from the speaker.
This back generated current can also cross-contaminate between the two channels. Having four wires all the way back to the plug dramatically reduces this contamination.
In a balanced setup, there is no common ground at all in the left and right audio circuits so separation is maximised and interference between channels is minimised.
The other big advantage of a balanced setup is that the two wires carry a mirror image of the signal but 180 degrees out of phase with each other. Any interference which manages to penetrate the cables will be in phase on both + and - wires and will thus present as zero volts across the driver. This is why balanced connections were invented - so that signals could be sent over long distances while being immune to interference.
I hope that's right. If not I blame alcoholic intoxication (well it is Friday night, hic!)