Look frens.
I'm not convinced it's anything other than Optical Character Recognition (OCR) errors.
Many of you have scanned text and performed an OCR operation on it. Some using free software or crappy software that came with your scanner, or maybe you bought a high-end Adobe package.
Besides text, I have also scanned Sheet Music and seen how the OCR functions in doing a poor job of recognizing characters (notes, sharps, flats) that are too close together, or where there's some confusing xerox copier dots or smears or handwriting or slurs or other musical markings in the same general area. Better software does a better job but a lot of notes have to be corrected by hand.
OCR failure on sheet music just illustrates the general principle that software is not yet perfect at converting rendered or printed text into ASCII.
I don't have any specimens handy on this device to illustrate this but it's definitely a thang.
Also remember the dates when some of these documents were created and maybe subsequently scanned?
Software has gotten better but consumer-grade OCR software started to become available around, I think, 2005ish and was not very good. The software in use then is not nearly as good as what's available now.