Exploring the limits of science and beyond: research strategy and status

Journal of Parapsychology, The, March, 1994 by J.E. Kennedy

This internal pattern is evidence that psi is responding to, or operating on, the majority-vote outcomes rather than being distributed throughout the raw trials as expected. The pattern is evidence for some type of goal-oriented psi effect even if it does not actually reflect efficient psi operation. This pattern of psi effects focusing on the majority votes with low scoring rates on the raw trials has been found only in studies that specifically used majority votes to increase psi accuracy, and it has occurred in the great majority, if not all, of those studies. The pattern has been found in cases when the experimenters did not expect it and were at a loss to explain it (e.g., Brier & Tyminski, 1970), as well as in cases when the experimenters predicted it (e.g., Radin, 1990-1991).

It is unlikely that majority-vote and other communication theory methods can be used to increase psi accuracy or reliability if the goal-oriented psi hypothesis is correct. The experimental outcome or majority-vote outcome becomes the goal rather than the individual events comprising the majority vote, as assumed in communications theory. The study by Schmidt (1974) directly demonstrated this point. Suggestions for study designs that might minimize or circumvent this limitation, as well as some ideas on the interesting implications of efficient psi operating in a shifting hierarchy of goals, have been discussed previously (Kennedy, 1978, 1979).

Evaluating Simplicity

The goal-oriented experimenter-effect hypothesis is a good example of why simplicity must be measured by ease of testing rather than just conceptual simplicity. Conceptually the goal-oriented experimenter-effect hypothesis is extremely simple. Many experimental parameters (sample size, subject selection, etc.) that are important under the traditional assumptions for psi research are dropped from the model. However, the goal-oriented experimenter-effect hypothesis makes the interpretation of research much more difficult and is therefore a more complex hypothesis to investigate.

Beyond the Limit of Testable Hypotheses

The preceding discussion of simplicity in science indicates that many important hypotheses are outside the domain of convincing scientific tests. Hypotheses pertaining to life after death and religious beliefs are obvious examples. Although the public widely holds religious beliefs and believes in life after death, these concepts are basically ignored in scientific research, while simpler, testable hypotheses are considered. Of course there is no empirical evidence that these untestable concepts are actually incorrect. Because scientific progress conceivably could lead to direct tests in the future, these concepts are more properly classified as deferred or "on hold" rather than wrong.

What role, if any, does science have for concepts that are widely held but are beyond the present limit of testable hypotheses? Is science a basis for dismissing such concepts as unscientific, irrational, and indicative of maladjustment? Alternatively, is science impotent with regard to these topics? A variety of factors bear on these questions.