If I read the file as .utf8, AND strip out the UTF16 BOM, BINGO.
let cleanedClippings = clippings.replacingOccurrences(of: "\u{feff}", with: "”)
Each entry has a UTF16 marke? who knows. 332 of them in my file! I grabbed it off my kindle again just to see if I had somehow done it, but no that's how it is.