From what I've read 3-wire DC was already an attempt to cut voltage drop-related losses in local-scale DC distribution networks spanning roughly the size of a neighbourhood. The first DC systems were plain 2-wire. In the Budapest Electrotechnical Museum you can see the very first working transformer, which is a symmetrical three-phase transformer so I'm fairly sure 1-phase 3-wire AC came later but the question is how much later.
120 V AC never made an awful lot of sense TBH. The voltage is too high to be considered safe (the limit is 50 V AC or 65 at the highest) but too low to drive substantial loads without requiring fairly high currents and large conductors. Higher currents also cause higher voltage drops. Just to give you an example: a cooker (range) with three sets of elements each connected to one phase and neutral of a 230/400 V system could be connected using 14-4 NM cable using a 3-pole 15 amp MCB. The same load connected to 230 V requires a 50 amp MCB if full functionality is required (although most European countries factor in load diversity as it's unlikely that all elements will draw full current at the same time as the elements are thermostat-controlled) and allow for 25 or 32 amps to be used).
Mains voltage is always a compromise between insulation and low currents. Frequency is mainly a compromise between equipment size (transformer cores, motors and generators) and motor/generator RPMs, the higher the frequency the more poles a slow-running motor or generator needs to have. I think we'll likely be stuck with 50 and 60 Hz forever since changing a large grid's frequency seems pretty much impossible unless the grid is split up into smaller sections temporarily DC-coupled and since more and more equipment accepts more than one frequency I can hardly see such a conversion being feasible either. Voltages on the other hand may change, as South Korea has shown. As far as I know the country changed over from 120/240 V 3w to 230/400 V 4w in a period of only two or three decades (1970s to 90s).
