In 1986, I argued that pains are essentially not phenomenal states. Using a Wittgen-steinian son of argument, I showed that the same sort of phenomena can be had on different occasions, and on one occasion persons be in pain, while on another occasion persons not be in pain. I also showed that very different phenomena could be experienced and, yet, organisms have the same sort of pain. I supported my arguments with empirical data from both laboratory and clinical studies. There (...) is nothing about this thesis I would now retract. However, there was a further thesis that needs to be reconsidered. I argued that phenomenal states are only accompaniments of pains, that pains are essentially a combination of cognitive, affective and behavioural/motivational states. This thesis I do now zaish to retract. I now argue that phenomenal states are necessary for pains, but still not sufficient. There must also be a cognitive state which involves an evaluation of the phenomenon as something like, 'Harm to the body'. The evaluation is a kind of de re belief, regarding the phenomenon as itself representing harm to the body. Besides admitting that phenomenal states are necessary for pains, I also now claim that other relevant belief states, affective states, and behavioural/motivational states are not necessary for pain, but normal consequences of pain. This revised theory is preferable to the 1986 one because it fits better with empirical facts (including providing better explanations for anomalous cases), fits better with certain powerful common-sense intuitions, and fits better with a larger theory of consciousness I have been developing. Among other things, it turns out that being in pain is a quite peculiar conscious state and considering it as a paradigm for consciousness is a serious mistake. (shrink)