Actually it doesn't. That's the problem with granularity, you're chasing the difference between state (which is never defined at a single point for a switch) and scan time of the input system. If one switch is triggered within 1ms from the other, they have been triggered simultaneously, with any resulting error being the result of contact bounce. For a system that has a scan time of 1 second, it could very well have been that one switch had contacts which hadn't settled one side of the second mark while the other had to wait for the system to process its input logic again 1 second later. That's the OP's point. I don't think it makes any difference to the outcome or investigation, but it is important in some situations.
I have a real world example of this which I have experienced. On a Triconex industrial safety system there is a single switch that sets the system to OFF, RUN, REMOTE, and PROGRAM mode. It spends most of it's time in RUN, but I had to download some new code to it so I went to the system and flicked this switch to RUN. Unluckily as I flicked this single switch 2 of the system's 3 main processors saw one contact close 125ms after the 3rd processor, that was enough to trigger a diagnostic error and force the processor to reboot raising all hell in the control room as system fault alarms were coming up. One switch only, two different states in the same system.