The only place I've seen a clearly defined reaction time criterion is in the critical engine failure case for multi-engined aircraft in JAR/FAR-23, there's no such reference in JAR/FAR-25. My feeling is that this is probably because the case is far more critical in a lighter aircraft.
JAR/FAR-23 requires that with flaps and gear up, MCP and propellers set for climb, it must be possible to fail the critical engine at any speed down to Vmc, wait 2 seconds (the defined reaction time), and then bring the aircraft back under control without seeing more than 45° of bank.
JAR/FAR-25 don't define a reaction time, presumably leaving it to the test team to decide what is appropriate, but are stricter in requiring a maximum of 5° of bank and a 20° heading change without using more than 150 lbf rudder force.
However, going back to JF's comments, there are all sorts of cases (single engine aircraft engine failure in a steep climb for example) where it is necessary to determine a pilot reaction time and use it during certification. In my experience 2s is pretty much universal in civil certification, presumably because the figure already exists in one context.
If you go back 30 years and trawl through either AvP970 or BCAR Section D, which I image were set texts when JF did his ETPS course, you'll find that they were much more prescriptive about such things than modern practice, and you'll probably find a lot more wisdom there. On the other hand, BDN has a long track record of not even reading the standard and simply applying common sense.
Another source is Donald McClean's book "Automatic Flight Control Systems", which defines response control laws for pilots at different states of ability and alertness. Where he got them from I don't know, but he's still a professor at Southampton University, and I imagine would be glad to talk about it if you phoned him.
G
[This message has been edited by Genghis the Engineer (edited 29 April 2001).]