Arcing is caused by the electrical inductance of the supply system
Isn't arcing caused by the breaking of a circuit carrying a (sufficiently heavy) current? I know inductance in the circuit can help it start, but for example you can do arc welding with car batteries and a welding rod.
The term 'spark' is often used for a brief luminous discharge e.g. at switch contacts when a current carrying inductive circuit is broken, and 'arc' when the energy available is enough for the spark to evolve into an arc and conditions are right (e.g. the gap has not widened too much).
One thing that surprises me about dc units is that they only have pickups on the bogies at the ends of the units (the bogies under the driver's cab). It must give an increased risk of a unit getting stranded.
There is a "train line" running through the whole unit linking all the shoes to reduce the chance of gapping. Older Southern units prior to 1951 stock (EPBs etc) had power jumpers on the ends so that coupled sets could have all shoes linked. I believe that only depot staff could connect/disconnect these. Underground stock does not have a continuous train line due to DOT regulations for running in lines that are mainly tunnel, this includes 313s in DC mode. (Reason: some pretty bad historical fires caused by arcing)