In general, if both systems are mechanical stabilization and are trying to correct the same camera movement independently, they can overcompensate rather than help. That’s why the usual expectation is that one system should be disabled, or that the camera would need to coordinate which system is active.

A possible exception is when the two systems correct different things, or when body stabilization is digital/image-based rather than mechanical. In that case, the body might only correct residual movement the lens did not handle.

So the practical answer is: unless the camera and lens are specifically designed to communicate and coordinate stabilization, don’t assume they will stack usefully. With uncoordinated mechanical lens IS plus in-body stabilization, overcorrection is a real risk, and disabling one system is the safer approach.