I often see my meter indicate 5 -6 volts of AC with the leads in the air. It has a high impedance, and there's lots of 60 Hz around, especially under the fluorescents.
I can't quite envision how you shorted the outlet and still had voltage indicated on the meter. Was that while the short was in place. Be careful about doing stuff like that - the voltage might be as the result of a wiring fault and the resulting arc could hurt you, plus potentially cause a fire.
With your voltmeter NOT connected to anything, short out the inputs to your voltmeter and ensure that it reads zero volts. Most people will say it isn't necessary to short the inputs and I would agree. But my point is, if your voltage reading with the leads shorted is different than your reading with the leads not shorted, then the meter is picking up a signal, possibly from the flourescents.
Very funny, although I'm glad I'm getting forum recognition.
The DMM works fine still, the volts it is reading aren't the "phantom" reading that I get when the leads aren't connected. The reading will usually be a few mV unconnected then I will test the outlet and get 17V.
Perhaps the fluorescent light is related. I turned off all the lights in my house and the reading dropped from 17V to 16V. When I turn the light back on the voltage goes back up to 17V again.
I tried connecting a 340ohm resistor and multimeter in a series an got a reading of 7.5uA so I guess wherever this voltage is coming from it can't hold a charge for very long
i didnt get the experiment from the op , by any chance are you including the neutral in that measurement? mostly the loads are not balanced for a three phase network and the neutral is not always at 0 Volts as a result. (my friend measured it to be around 12 V in his house.)