(Nepster's proposed fix for pre-placed blockers: When two pre-placed blockers overlap, the blocker with the higher ID walks).
This could be generalised to "When two blockers overlap, the one that's been a blocker for the least amount of time walks" - if we see it as that, since the lower-ID preplaced lemmings get placed first, they've been a blocker for a nanosecond longer.
And thus, this rule could then be applied to if this situation arises mid-game as well. It could lead to some messy results, but it does provide a consistent rule that absolutely prevents blocker overlap, and leaves teleport-blocker allowed. Possible implementation: Have a variable that keeps track of which frame a blocker was assigned to the lemming on. In the case of two overlapping blocker fields, the later-assigned one walks.
I can think of one way such a mechanic could be abused to create a level with a very obscure solution, but I would think this situation would be
so difficult to set up that it probably wouldn't happen in practice outside of a test case, or a level that makes it very obvious that the trick is being used.
The one other solution I could think of that would still be fairly consistent, is "treat teleporters the same way as exits, in terms of what can or can't go through them" - IIRC, blockers cannot exit (whereas other skills you'd generally want to pass through a teleporter, such as a builder or basher, can). Instead of creating a new special case, this simply applies an existing one to another situation, and this somewhat makes sense as most teleporter designs convey the impression that, in-universe, a lemming would need to actively choose to enter it, just like they would with an exit. This will, however, break any level that requires on blocker teleporting (or perhaps faller / floater / glider / falling-ohnoer teleporting). Nevermind, seems blockers
can exit, so this wouldn't be consistent.