1. John North suggests the lintels created artificial horizons over and between which stars, moon and sun were ritually observed rising and setting, the observation points being from ditch-henge and avenue. Who knows?

2. I have, without having seen anyone else's suggestion, also pointed out that the "mortice and tenons" are not. Can't possibly remember what thread that was in. So yes, they don't provide the anti-twisting functions of true m&t's. But what they, and the tongue and groove joints DO provide is the classic just-good-enough-therefore-very-efficient function of "stopping the starting."

No idea what a real engineer would call this. But I know from experience that it's often not necessary to prop, rope or otherwise constrain a load against the worse case slippage or tilt. Most of the time, you merely need to constrain it against the beginning of the movement, the smaller average jostle. This will be enough to prevent sympathetic vibrations (swings, slippage, etc.) from building up and reaching a catastrophic tipping point.

In the case of the lintels, you obviously need to keep them centered over the uprights. What happens, though, is that without joinery, a little bit of creep --just from, say, frost heave-- begins to create a little bit of lean. Which then allows a bit more creep. Which creates a bit more lean, and eventually the two movements exacerbate each other to catastrophe.

The original Stonehengineers weren't trying to create joints which would resist a sudden strong force, such as a big lorry smashing into an upright. For that, yes, you'd need a true, all-the-way-through m&t. They were, quite rightly, using the simplest kind of joints that would resist the small movements they expected would otherwise occur.

3. Yes, many menhirs continue to stand. As do many dolmens. But how many, compared to how many were erected? Plenty of both have collapsed. I doubt we have any good idea of the ratio or proportion of collapse to endurance. And few if any other ancient structures are the Stonehenge free standing upright and lintel type, in which a pretty good portion of the structure's weight is on top, potentially sliding slowly off a single center. Menhirs have no tops. Dolmens have triangular or quadrilateral polygonal base points, not double, linear base points.

Just a few thoughts on simple explanations for a seemingly complex phenomenon!