I studied psychology and business in school so I'm not an expert on anything history related but you should know that SLA Marshall and Dave Grossman are not considered credible and Grossman's "only sheepdogs and sociopaths are likely to kill" theory doesn't hold up- there's mountains of evidence that most normal people will attempt to kill their enemy. The question raises way more holes than answers. And the phenomenons he claims his theory explains (like elite/trained forces being way more lethal) are explainable many other ways. Crime statistics make no sense based upon his theory, much less his "final conclusion" in the end of the book that video games will train people to overcome their reluctance to kill and therefore will lead to a massive increase in violent crime. Grossman made this prediction in the 90's...

The SLA Marshall thesis is highly disputed. In particular, it's possible that his "painstaking research" never existed and he just made it up because it sounded right.

That said, "giving someone pardon" in early modern European combat was very much a thing; like people had done in the Middle Ages, people would take their pardoned enemies prisoner for ransom. Armies would also take prisoners and assimilate them into their own ranks. Finally, I have seen in a letter to a soldier in the 1620s from (I think) a family member of his, mentioning that their neighbor had joined the opposite side, and saying "If you see him on the field, give him pardon."

So it definitely happened, but out of an instinctive reluctance to kill? Western European Christian society during the period that I study is far more violent than our own culture. I don't know how much is "innate" and how much is "taught" but I suspect that they would be far less likely to fail to kill someone on purpose than we would.