Very interesting, another phenomenon I have not observed before. Since the discharge persists for so long and there is no frictional force to create a sustained electrostatic recharging, I think the mechanism of this light generation in your lamp is different.
Possibly it is caused by a ghost voltage on the power supply to the lamp, which exists even after the switch is turned off. This is very common in many modern buildings, where electrical cables are routed parallel to each other over lengths of many metres. If the switch to a lamp socket is turned off but that cable is running alongside another live cable, its not unusual to measure voltages above 100V at the lampholder due to indiction from the adjacent live cable (or course since there is no direct connection the current flow is miniscule). However if you touch the lamp, you then complete the circuit to ground via the discharge tube and the few microamps current flow can result in light generation.
In the beginning of the CFL ays (and again with LEDs) this was a big problem, especially in the wiring of modern hotelrooms with many lightswitches and cables adjacent to each other. In the early lamps the leakage current would cause a capacitor in the driver to slowly charge up over some minutes, and then repeatedly discharge with a bright flash from the lamp or LEDs. Eventually lampmakers included a tiny ceramic capacitor and high value resistor connected directly across the mains input of the driver to prevent this effect. Since about the last 5-10yrs most retrofit lamps have this ghost voltage filter, but some of the cheaper or smaller lamps do not and then we can see an effect like in your video. One sure way to test this would be to unscrew the lamp from the socket and see of it still lights.