In reply to barnaby:
It's easier to think about this in DC, but the physics work in AC too. For simplicity, let's say we have a generator of fixed power output (which we can then transform, losslessly), a long transmission line, a load of fixed power requirement which matches the generator output (with a lossless transformer before it), and a ground return line.
That's a misinterpretation of Ohm's Law, and a common one. V=IR in that case refers to the
voltage drop across a component, not the 'voltage through' it (see below). Across a long transmission cable, the aim is to reduce the current, and hence the voltage drop (which by P=VI is proportional to the power loss in the cable). The aim is to have the same nominal voltage at the other end that we did at the start.
The voltage of the outbound cable's positive end above ground potential only becomes relevant when we consider the whole system - transmission line, loads and ground lines. Increasing this using transformers will not increase the current, because the load only demands so much power, and the generator can only supply so much. Hence, increasing the voltage decreases the current, which decreases the voltage drop across the lines, and hence the resistance losses in the system.