This is just one example, but it shows how what is a fact is determined by what assumptions one uses to interpret data. If one assumes a linear time-line of causality (which imo is fading as scientifically valid), then it is a fact that death preceded man perhaps. If one assumes the system can be affected as a whole, and that non-linear causality is possible, then it is not a fact that death preceded the creation of man.
Indeed, the assumption 'assuming time is linear' is an assumption.
Should we concerned about this? Well I don't think so, but some people like yourself are. If science is to seriously consider that their assumption is questionable, we must search for some evidence of this reality changing entity. Some evidence that things such as 'death' have been added to the program. Your concept sounds pretty unfalsifiable as it stands. After all, why would there be any evidence? If all of history is changed, then all the evidence changes to, so we are left with nothing.
As such, it sounds like a philosophical objection, which is entertaining its possibilities, but not something we should seriously consider in our daily lives.
I could say that there is a more important assumption, that outweighs yours in scope by several orders of magnitude. We assume that reality as we perceive it is reality as it is. We could just be 'minds/brain' in a 'jar' being fed 'sensory' information simulating an alternative reality with whole new rules. Rules which we have to figure out - but which we are given a limited amount of time to explore. Any of our observations therefore could be part of this 'game/experiment/coming of age ritual/whatever' and the rules could change at any time, including making those rules retrospective.
In your scenario we could have a fake reality where death doesn't exist (maybe death doesn't exist in reality), but people grew bored of staying alive and they introduced a death algorithm to the fake reality...
We could make Cartesian type objections on our assumptions such as this till the cows came home, it's fun and its a great way to develop interesting sci-fi (eg The Matrix, Dark City). It's not, however, a really great objection to scientific theory and perceived facts based on the evidence.
Can you imagine it? "No your honour, I didn't commit the murder! The evidence was added to the system", or "The fact that I committed murder was added to the system, retrospectively. I had no control over this process so I cannot be held morally responsible".
I would submit on every thread, the underlying assumptions ought to be fair game to be challenged as these assumptions determine what is a fact and what is not.
I disagree that it is practical to do this - since any assumption can be challenged by any party at any time. Threads would end up getting bogged down in philosophical quagmires. Such asides should be raised as asides, and if they start to become a focus point of the topic a new topic should be proposed. Can you imagine a topic that was about the nature of the fall getting bogged down with 'you are assuming the bible supports a fall happening' and 'you are assuming the bible is inerrant'? Topics would quickly become boring and unfocused - indeed we see that this happens often enough as it is, and mods have to step in as asides become tangential marshland.
Topics cannot do with this kind of thing. That an objection exists can be raised, but should rarely be the central focus of the topic unless the objection is reasonable and directly related to the topic at hand.