After a hundred years, psychical research (and experimental parapsychology) are still regarded as pseudosciences. Despite sceptical attacks, however, work continues: the results cannot be lightly dismissed. Evaluation of this work demands philosophical attention, since there are large metaphysical assumptions made, and the experimental results are vulnerable to logical criticism. The supposed kinds of paranormal phenomena are anatomized, and the current terminology criticized. Broad's basic limiting principles are examined, and the extent to which psychical research might threaten conventional science is considered. The metaphysical foundations of psychical research are examined. Psychical research takes a dualist view of human beings, although its experiments are behavioristic. Survival of death is assumed to be a matter susceptible to empirical verification. These assumptions are critically examined. I argue that ESP neither gives us knowledge, nor is it a form of perception. Although psychical researchers interpret their results in causal terms, this is illegitimate. Backwards causation, therefore, need not be canvassed as an explanation. As the basis for a scientific discipline, spontaneous cases are inadequate, as they involve underdescription (lack of detail) and overdescription (tendentious exaggeration). Observational cases, given the absence of theoretical explanations, and. the prevalence of fraud, also fail as science. Hume's arguments on testimony, and modern versions of them, are considered. Experimental parapsychology fails because ESP and PK seem to occur without conscious awareness or significant physiological signs, and in a way that transcends space and time. There is thus no way that science can gain a grip on them. Without a theory and a mechanism, there is nothing but a series of statistical quirks. The proposed models do not mesh in with the experimental results. Paranormal forces, if they existed, might contaminate all scientific work, and no one could ever know. Taken on its own terms, psychical research is impotent to undertake the task it has set itself.