2.2V, being nearly halfway between 0V and 5V is often indicative that there is a short to that line from somewhere.
If it's build on vero-board/strip-board, check for slithers of solder, wire-strands and solder bridges. Running a sturdy craft knife between all copper tracks will clear many shorts, and any 'kick' or jamming may indicate where a fault is.
If it's a breadboard circuit, check that wires haven't fallen over or are slighly pulled out and are shorting one another near their ends. Check that the wires really do go where expected.
If the 18X fails and an 18A or another 18X works now ( ignore earlier tests ), then it would probably be fair to say the chip itself was damaged, but the process of swapping chips can sometimes cause the fault to appear or go away, but <b>firstly </b> ...
Check for any shorts there may be. If the chip in place has been damaged and the fault remains, you don't want to damage more chips by plugging them in.
Unplug the 'faulty chip' and check what the voltage is at that pin on the socket; if it's around 2.2V you have a problem on the board.
With the power off, do a resistance check between the socket pin and 0V, 5V and all other points to which the track from that point connects.
You can then try with the PICAXE in the socket but having gently raised the leg horizontal with a pair of pliars so all but that leg is in the socket. Check the voltage on the leg; if it's 2.2V then it would seem that the PICAXE has been damaged. Otherwise, check the voltage changes as expected when the button is pushed
You can straighten the leg and plug the chip back in and measure the voltage on the pin. If it's not 2.2V the fault is intermittent or has cured itself.