Duke, I checked the throttle linkage microswitch, and it's action seems consistent. I did not check the TPS - that much disassembly will require more time than I'll have for awhile unless the car becomes undriveable.
However, I didn't think these would be the problem. My reasoning is this: without touching the accelerator, the idle problem ultimately responds only to one thing I can pinpoint - turning the engine off then back on, which should not move the throttle linkage. It specifically does NOT respond to 'blipping' the throttle linkage. My reasoning could be faulty - current surges through switches could be temporarily cleaning contacts, or other such weird issues. Still, I have put this possibility below the prime suspects.
Neil, I did check the pot, and behavior is less than ideal at small displacements. My ohmmeter shows that contact may be poor, but not to the extent that it goes 'open'. Still, I would think that a high idle would be a symptom of too much airflow, and the flow meter should only affect mixture. I know that with other rpm monitoring idle systems you can adjust the mixture over a wide range without affecting idle speed, and my impression is that is true too of these CIS systems.
Too much airflow points me to the idle air control valve, but I had cleaned it previously, and the problem that corrected (rough idle and occasional stalling) is quite different from the current symptoms. It also seems unlikely turning the engine off then on would be the only thing that affects the problem if it were this device, or any other air 'leak'.
Also, I am factoring in the fact that after extensive futzing and replacing parts, the idle had returned to normal - for a month or so. The last thing I replaced was the OVP, and the symptoms now tell me I have a confused ECU. It is getting an incorrect signal for engine rpm OR it is trying to control engine speed but can't (air leak, sticky idle air valve) OR the brain is scrambled (bad new OVP or bad ECU). Am I offbase in this line of reasoning?
Thanks to those who have responded, and I WILL eventually be checking more into the airflow meter and throttle position sensor, even if they are not the current problem.