The CT's position, however, does project a certain aura of empiricism, and no sane philosophy of science, even a hyper-rationalist one, can completely exclude empirical evidence as being irrelevant. But I think a careful examination of the remnants of empiricism that do exist in the CTs' program will serve primarily to further illustrate the dominant status that the coherence principle has achieved in their thinking.
CTs do claim that they would accept "the existence of psi" if conclusive empirical evidence were forthcoming. However, for such evidence to qualify, all conventional explanations must be completely eliminated, even if they are unlikely or implausible. This is the famous position taken by C. E. M. Hansel . I have never known any CT to repudiate it, although Ray Hyman has taken a few bites around the edges. The explanations proposed by CTs often are implausible and are not even defended as being plausible. Most assume dishonesty or gross incompetence on the part of the experimenter in carrying out elementary research procedures. If all else fails, the CT stalls for time, suggesting that a flaw in the procedure may be revealed at some later date. This of course is true, and dishonesty and incompetence could indeed account for all the experimental evidence; there are precedents for both. However, none of this takes away from the extreme elasticity evidenced in the CT approach; one never really gets a clear sense of what CTs would minimally accept as evidence for paranormality. Exactly how elastic the CT approach is in practice will only be known as the empirical evidence becomes harder and harder to explain away, but it clearly can be carried much further than it presently has been. Sooner or later it can be overwhelmed by data, but the CTs have made this as difficult as possible to achieve. The problem, in a nutshell, is that the coherence principle is used to arbitrate what counts as acceptable empirical evidence.
Not only does the existence of an alternative explanation render a psi experiment worthless as evidence for paranormality, but many CTs take it to mean that there is no reason to even question that whatever happened has an adequate conventional explanation. Event though CTs will often deny that any one of their specific alternative explanations is necessarily what happened, at the same time they will insist that is nothing to explain. This conclusion can be shown to imply that the correct interpretation, whatever it is, is conventional. The conventional explanations are not necessarily accepted individually, but they are accepted as a class. What transforms a set of possible explanations into a set of acceptable explanations is—you guessed it—the coherence principle.
To see how this follows, consider a statement made by Ray Hyman in his critique of the ESP and PK experiments by Helmut Schmidt using automated methodology. This research is widely regarded by both parapsychologists and CTs as among the best the parapsychologists have to offer. The quote is as follows:
Only when the parapsychologists settle upon a standardized paradigm, tidy up the procedures, demonstrate that the results follow certain laws under specified conditions, and that these results can be duplicated in independent laboratories, will we have something that needs "explaining." (p. 39)
Although I take exception to some of the premises of this statement, in particular the one about the lack of independent replication, I wish to focus on the conclusion of the statement. The logical equivalent of this concluding phrase is that Schmidt's results constitute nothing that needs to be explained.
Now up to the time that Hyman's paper was published, Schmidt had published 22 separate experimental series. Of these, 20 yielded statistically significant deviations, and 18 were in the predicted direction. Obviously, something is happening. Now either we have an adequate explanation of this something or we don't. If we don't, then it follows that we do have something to explain. Thus the only remaining way to interpret Hyman's statement literally is that we already have an adequate explanation of Schmidt's results. Even if we take the statement in the looser sense to mean that Schmidt's results provide nothing that scientists should bother themselves about (which I suspect is what Hyman really meant), the implication remains that the population of potential explanations includes nothing very interesting. Since a paranormal explanation would be quite interesting, even to a CT, I think the reader is entitled to infer that Hyman does not take the possiblity of such an explanation seriously.
It is true that CTs occasionally attempt to provide empirical support for conventional explanations of OPEs. Most often this is in the form of what they call "debunkings", exposés of particular psychic claims or psi experiments. The most succesful example I can think of is Scott and Haskell's and Markwick's detection of irregularities in the data of the S. G. Soal ESP experiments (Dr. Scott is one hard-line critic whose writings do reflect an empiricist value system, but I consider him an anomaly in this regard).
Taken as a whole, however, the CTs' empirical research program is not impressive by normal scientific standards. Its most damning feature is that debunking need not confirm that a conventional process occurred, but only that it could have occurred. Only a small percentage of debunkings actually provide empirical evidence for conventional hypotheses. Because of the absence of any systematic sampling procedures, it is even more unclear than usual how far one can generalize the "real" debunkings that are achieved. Debunkings are highly concentrated on publicity-seeking self-proclamied "psychics," generally one of the last places I as a parapsychologist look for "real" psi.
The preceding critique says more about the inadequacies of debunking as a research strategy than it does about how specific debunkings have been carried out. The greatest deficiency of the CT research program is it almost total lack of systematic research and model-building for assessing relevant conventional hypotheses on a broad scale (one exception is a handful of studies exploring the cognitive styles of "believers" and "skeptics" , but this approach is very much in its infancy). CTs often like to cite the research literature in general psychology concerning human capacities to misperceive or misinterpret events in line with preconceived biases, but such research is far from consistent in its implications and is not directly relevant enough to the anomalies at issue to be satisfactory. The premise that people's perceptions and cognitions are sometimes biased does not lead to the conclusion that such biases are responsible for the critical masse of OPEs; this is an empirical question. What is needed is research that is both (a) nomothetic and (b) targeted as closely as possible to potentially evidential psychic events.
This of course is a very difficult area, and CTs, were they to launch such a program in earnest, would run into many of the same difficulties (e.g. the elusiveness of the phenomena) that parapsychologists have had to put up with for decades. But no one has said that science is easy, and these difficulties cannot be used as excuses to justify basing conclusions on insufficient evidence. If CTs were to apply the same ingenuity to serious research as they have to some of their more celebrated debunkings, they might be able to come up with some sound, broadly based evidence for their hypotheses that is worthy of being taken seriously by the scientific community at large.