First- I calculate replacement level by sorting every position by PPG/PIP, then finding the point where there were enough games/innings for every team in the league to reach caps. Those PPG/PIP baselines represent where that line is for each position, with MI/UTIL factored in. Effectively I want to assign dollar values to only those players that would have been worth starting, with the benefit of retrospective knowledge.
Second- Schechter’s book is really focused on 5x5, so when he talks about the non-catcher positions being indistinguishable, he’s really just referring to that format. I agree with him by the way, in 5x5 there is very little difference in replacement level between 1B and SS, for example. In linear weights scoring formats like ottoneu FGPts, which effectively scores based on wOBA, there is a much bigger difference in true hitting performance between a 1B and SS (though it is muted a bit by there being a MI in ottoneu, but no CI).