>>417
Thank you for this lovely VIP corpus. There seem to be some problems with whatever you used to parse it though, for example "quot" is very common probably because it interpreted the symbol " as """, and "don" is very common probably because it parsed "don't" as "don" and "t".