so what's wing rock? It's not tipstalling, in harrier you're basically flying the wing in the propwash, anything outside of that should be fully stalled.
I'd wager that wing rock is caused by the huge vortex that appears at the tips of the wing when it's at a high angle of attack.
I experienced *real* bad wing-rock when I used to build constant-chord unswept combat wings with tip-fins.
Here's how it works...
Any wing which meets the air at an angle of attack greater than zero degrees will have an area (usually a bubble) of low pressure on the top surface and higher pressure pushing on the bottom surface.
Naturally the low and high pressure areas would love to meet somewhere for a coffee and cancel each other out -- but there's a big fat wing in the way :-(
However, where the wing ends (ie: at the tip)_the high pressure air under the wing does tend to flow up and over -- to fill part of the low pressure area on the top-surface.
The result is a large spiralling vortex.
Now imagine for a minute that one wing produces slightly more lift than the other -- for whatever reason.
That wing will rise and the other will fall.
Now it becomes much *easier* for the high pressure air under the wing to rise up and over the wingtip because it's going to naturally move up-hill towards the tip.
This means that the tip-vortex will increase in magnitude, the pressure under that wing will fall, and the low-pressure area on top of that wing will reduce in size. The net effect is that the rising wing is now producing *less* lift than it was -- so it falls down and the other wing now is now generating more lift so the plane starts to roll the other way - then the whole process starts all over again so we get a rocking back and forth as each wing alternately creates more lift then spills it.
In normal flight, the size of the tip vortex is too small to produce any perceptable degree of wing-rock.
Slow a ship down and give the wing a huge angle of attack however and the yo-yoing tip vortices will produce a very significant oscillation.
There are ways to reduce this phenomenon -- but that's another story