You observe leaves shaking when there is wind (a correlation). To test for causality you need to make an intervention. You still the leaves - the wind keeps going. You block the wind - the leaves stop shaking. There clearly is an asymmetry here. The physical equations may be symmetric, but this doesn't justify calling a microphone "speaker".

>The physical equations may be symmetric, but this doesn't justify calling a microphone "speaker".

Actually a speaker is just a microphone with very small adaptations depending on the use (polarity, noise sensitivity, etc).

But this is not about the wind, but about sound. The wind you describe is external ("You still the leaves - the wind keeps going"). But you can just as well have a wind produced by the leafs themselves.