A "leadscrew" is similar to the hole position for an indexing plate. Each generation will decrease in error from the one previously. After the tenth or so generation, the error is so small that I can't count that small. Keeping in mind that each generation being taken from the one before has been taken before any significant "error" can be introduced into the master.
In the case of a lead screw, the error can come not just from the screw but also the follower(half nuts), the drive/driven gear train, and any number of other points on, for example a lathe, the master machine. There could be a speck of dust on the master that would introduce an error. But the error would be so small that a hobbyist would not have access to the instruments accurate enough to measure that error. Actually, at that level of measurement, the measuring instrument itself would have a questionable level of accuracy.
In the late '70s, I did calibration of an "optical pyrometer". A color match system where my color discrimination(eyeball) was a factor in the calibration process. My "master", the calibration stand, was
NBS (now NIST) tracable. It was calibrated by Leeds & Northrup in their Pennsylvania shop, which was in turn calibrated from the NBS reference. Basically third generation removed from NBS.
There is a similar process involved for mechanical measurements, the level of accuracy being such that there is tracability back to some standard that is so accurate that a hobbyist cannot measure any error. Someone, somewhere could, but that tracability would cost so much it would demand a good sized company to justify it.
The original post was on flatness(two dimensions), which is a dirivitive of straightness.(one dimension) There is no theoretical flat in day to day use. A surface block may be accurare to a millionth of a meter(a micron?) But it wouldn't be truely flat, just acccurate to that dimension.
Going too deep into my memory, and bringing up things I haven't remembered in thirty years. One step removed from preaching mode, I guess. So, I'm out of here...
Edit:
Found this video on resurfacing a surface plate. Thought it might fit into the original discussion.
.