The reason that they start with 4mA is for "loss of signal" monitoring. If the PLC sees 4mA is knows that it is at zero value, but if it sees 0mA it knows the signal has been lost and can generate an error.
Also in many cases, the 4mA is used to power the sensor.
I believe that the reason for "4-20mA" is that that is what the first major product of this sort used, and everybody after that just followed suit because there were "existing products we can leverage." And the reason that the first product did it this way is because the two engineers who were designing it decided over lunch that it would be a really cool way to do things.
As an engineer, I often see things done this way.
[This message has been edited by SolarPowered (edited 02-07-2007).]