As my categorically final contribution to this thread, I offer this link to a paper I read online in html form, but couldn't relocate until now as it seems to have moved countries.

If you can see the parallels (sic) between the main topic of controversy in this thread, and the phrases "out of order execution", "branch prediction" and "non-determanism", and especially if you see the logic embodied in section 2.3 "Multi-threaded ISAs", and would like to discuss that seriously, then feel free to contact me via the email on my homepage which will stay active until it starts getting spam.