The phenomenon here is called "inductive kick". It is what gets a 12 volt auto battery to deliver tens of thousands of volts to a spark plug.
While it's true that a transformer will not pass DC, at the moment of connection (or disconnection) DC is not DC. It is a changing voltage of close to infinite frequency (O to 48 volts in close to zero time).
The transformer has a magnetic field built up during the connection and a collapsing field during disconnection. This field flux induces a momentary voltage spike the secondary. That is, after all, what transformers do.
In a properly wired ribbon mic - cable - preamp input system, not only does the primary see phantom on both legs simultaneously, so no differential and no current flow, also there is no path to ground so again no current flows and therefore no inductive kick. If the transformer or mic cable is miswired or shorted in such a way that one leg of the system is grounded problems can arise.
Lots of studios have been using ribbon mics with consoles that do not have per-channel phantom for decades. The phantom is always on. Lots of Neves are like that. That's why some people will tell you that phantom killing a ribbon is impossible. Even though I know it's not likely to be a problem, I still triple check to be sure it's off before pluging a cable into my R*yer. It can be a very expensive way to discover a bad cable.
Depending on the turns ratio of the transformer and the magnetic characteristics of the core, a very large momentary voltage can be generated. That's why bluebird could shock his friends holding the primary of his 12 volt radio shack transformer with a 1.5 volt battery on the secondary.
I've destroyed an SM57 with phantom and a bad cable. It sounded like a kernel of popcorn popping when the diaphragm tried to jump out of the mic case.