While it’s certainly possible that Pavlov ended up ‘doing’ operant conditioning on himself, I think something different is going on here. I think Pavlov is a victim of humanity’s language capabilities: we do not only link concepts on one direction (dogs -> bell) but do so bidirectionally (dogs <-> bell). There is a whole line of research that goes into how associations (which are the ones in classical and operant conditioning) are not the same as relations (which are the ones that we humans are proficient at, blessed with, and cursed with).