If the what becomes stale, you can tell. If the why becomes stale (and it can become stale), you'll never know, unless the what is also included.

The why becoming stale is a feature, that's when you know there is a VALID reason you thought this looked weird and convoluted, instead of you completely missing the inherent complexity of the problem.

The reason you want people to document the "why" is because you can easily check if one reason has become stale, but you can never check if every single possible reason is still valid.