On 13 Feb 2009, at 20:43, Stefan Monnier wrote:

>> Do we have a way to detect the modifier state from the Lisp level,  
>> or bind
>> commands to modifier state changes?
>
> No.
>
> I did some simple experiments with it, in order to get events like
> <control> and <down-control> (which actually were <down-control> and
> <C-control> because when the release event is received, the modifier
> state is not yet cleared).  I did get it to work OK without too much
> trouble (mostly I had to add some function-key-map entries to drop  
> those
> events when not bound, and I had to rework C-q so that it waits for
> a "real" key press rather than stopping at the <down-control>), but
> haven't investigated much further.

And I guess there's a bit of port-specific work involved, too.
It would be excellent to have this post-23.1.