Journal of Parapsychology, The: Correspondence

To the Editor:

I recently read Robin Taylor's review of Dogs That Know When Their Owners Are Coming Home (JP, 65, 90-96) and felt an immediate desire to defend the book and its author, Rupert Sheldrake. I turned to Sheldrake's response (JP, 65, 199-202) and Taylor's answer to Sheldrake (JP, 65, 202-204). Sheldrake presented a technical argument, but I still feel he did not cover the heart of the issue: intent. Who was his intended audience, and what were Rupert Sheldrake's intentions in writing the book?

Sheldrake wrote the book for the masses, and not just as an outreach. His target audience was also the general pet-owning public, some of whom were participatory and supportive throughout his investigations and writing.

As for scientists, the subject of pet-human consciousness interaction is one that professionals in all areas--including most parapsychologists-have neglected. Sheldrake has presented his research in journals, before professional audiences, and placed some of his texts on his Web site. With his findings available to scientists, the focus for his book was the public.

Sheidrake's intention in writing Dogs That Know When Their Owners Are Coming Home was to increase awareness among pet owners, students, and possibly others who could take part in further research. His goal could best be achieved by reaching the public, which is the issue I want to present from my perspective as an author of a popular book.

My original draft of Life After Loss was some 437 pages and read as if it was written for professionals in the field of bereavement. Raymond Moody, my coauthor, said, "The public won't read anything beyond a sixth-grade level." Four drafts later I thought I had deleted every hint of academia, but publishers still rejected it for being "too academic." It became a scant skeleton of the original, but my focus was to help the general population. Similarly, for Sheldrake to reach a wide range of readers, his book could not have been more technical.

Thus, I applaud him. What difference does it make if we fill our laboratories, offices, repositories, and heads with scientific discoveries if those findings never see the light of day? Who is being helped now, and who will that help after we are gone? Rupert Sheldrake's work has made, and will continue to make, a difference because he has brought it into the light. What an asset he is to the field of parapsychology!

DIANNE ARCANGEL

6831 Cedar Lawn

Pasadena, TX 77505, USA

dianearcangel@aol.com

To the Editor:

The question about the alleged elusiveness of psi phenomena is an important one and deserves extensive discussion. Kennedy's article (JP, 65, 219-245) is a good starting point, but there are a number of lacunae that should be pointed out.

First of all, not every psi researcher would agree with the position that psi is elusive and weak. For instance, Utts (1996) in her assessment of the Stargate remote viewing (RV) research claimed that this line of research shows great stability over time. May (1996; personal communication, June 28, 1996), who was responsible for a large part of this research, supported her assessment. The same can be said for the effect size, which, according to another review by Utts (1991), is not weak at all but can be characterized as moderate (in the order of 0.2).

Thus, when proposing a model, as Kennedy did, to account for the general elusiveness and weakness of psi phenomena, it would be a good idea to also take into account the apparent exceptions to the general rule (see also Bierman, 1980, about negative reliability).

I welcome Kennedy's position to view the elusive nature of psi not as an obstacle to overcome but as an intrinsic aspect of psi, although I am not sure why that would follow from a correspondence principle. One should realize, however, that in order to study (rather than to avoid) the elusiveness, it could be that repeating the same (boring) ganzfeld experiment is a good idea. The general attitude in the psi community, however, is that we should not stick to the same paradigm all the time. Changing a paradigm has even been proposed as away to "overcome" the elusiveness. Following this suggestion, the proponents of changing paradigms would never learn what the dynamics are behind the alleged elusiveness.

Kennedy discussed a number of hypotheses that would account for the observed unreliable and weak character of psi. These correspond largely and overlap to some degree with the hypotheses I put forward in my invited contribution at the PA convention in 2000 under the title "Descartes Error? The Nature of Psi and the Relation Between the Subjective and Objective World" (published, slightly modified, in The Physical Nature of Consciousness Bierman, 2001). In this contribution, which apparently escaped Kennedy's review, I discussed long-term drifts of effect sizes in several psi databases and concluded that, apparently, effect sizes can decline but also rebound over time. Skeptics would claim that this shows that we are unable to replicate studies: A significant decline implies nonreplication of an effect size, even if the overall database is still significant, as is the case for the ganzfeld database.

If psi researchers are asked about the meaning of the findings of parapsychology, they tend to respond that the reality of psi suggests that things are connected that we thought (based on classical physics notions) were not connected. This "connectedness" transcends space but, considering the evidence for precognition, also time. If we take that kind of generally claimed meaning seriously, and why shouldn't we take our own words seriously, then it follows that replicability cannot be expected. Replicability assumes and requires independence of experiments. But built into psi intrinsically is a connectedness that connects experiments with each other--not only in a trivial way, causally forward, by variables like increased experimenter boredom, but also in a way that we cannot even conceptualize, where (aspects of) future experiments in some way have a relation with current experiments.

This is not the place to comment on each of the hypotheses put forward by Kennedy. Most of these hypotheses are just "not crazy enough." I focus on the hypothesis that the observational models (Millar, 1978) might provide a framework to understand the elusive character of psi. And I would like to add von Lucadou's (1990) system theoretical approach to account for elusiveness. I will not discuss Palikari's balancing model, although it might also be a relevant framework.

Under the title "Psi Is Influenced by Many or All of the People Who Are Interested in the Potential Results," Kennedy discussed the observational theories as explaining elusiveness. He based his analysis on two sets of data: checker effects and effects on prerecorded targets (for a quantitative review of these, see Bierman, 1996).

It can be concluded from these data that

a. Experiments are not finished at the time that an experimental psychologist would say that the experiments are finished.

b. This time extends beyond the first observation of the results.

Kennedy incorrectly suggested that "many or all people who are interested" (p. 234) are involved. From the experimental results, this cannot be concluded. The observational theories come in many flavors (Millar, 1978), and only one claims that all future observers matter to the same degree. It has been shown mathematically that this leads to a divergence problem (Millar & Hartwell, 1979), and Kennedy referred to this problem by stating "the hypothesis of backward influence does not explain why the net integrated psi effect would make psi elusive rather than completely suppressed or enhanced" (p. 235). But, as argued above, such a complete suppression or enhancement only occurs if there is no limit to the number of relevant observers.

If we take a bit more abstract point of view by interpreting unreliability as a consequence of error variance due to uncontrolled variables, then it seems to me that the fact that we generally do not control future variables in our experiments could account easily for the apparent unreliability.

It is interesting that, within the framework of the observational theories, error variance ought to decrease in cases in which the future of the data is better controlled. This is where the apparent stability of the Stargate RV experiments enters the picture. These experiments were largely classified and followed a very strict protocol before the results could be released.

It can be shown easily by Monte Carlo methods that, given a mean effect size around 0.2 (averaged over widely varying individual outcomes of, let's say, 32 sessions), an associative RV experiment using the roulette as a random number generator determining the target would result nearly always in a riskless profit in the long run (using a safe betting scheme). The fact that since the end of the Stargate episode we have not seen the use of the claimed stable effect size in this way suggests to me that this effect size, once outside the closed environment of classified research, cannot be trusted to be as stable as inside this environment. This supports to some degree the observational theoretical "explanation" of elusiveness.

In his analysis of Complex Meaningful Information processing systems, von Lucadou (1994) stumbled on a remarkable similarity between the descriptive formalism and the well-known quantum formalism. The observational theories as discussed above had their roots in an analogy with quantum systems (Walker, 1975, 1985), while this new theory turned out to be similar to quantum theory without explicitly starting from it. In von Lucadou's systems theory, one gets psi as an equivalent of nonlocality in quantum theory. According to von Lucadou, there is no (classical) information transfer, but correlations arise. Once one tries to use the information, the correlations disappear, just as they disappear if one tries to use quantum nonlocality to transmit classical signals. In quantum systems, this remarkable effect occurs as soon as one creates the possibility of using the nonlocality for information transfer, even without actually doing so. This is very awkward and really looks like psi elusiveness.

That we still find correlations in psi experiments can be accounted for if one realizes that, in most cases (with some exceptions; see for example Carpenter, 1991), we do not manipulate the "source or target" information, but we let a random decision determine what the source information is. In this situation, meaningful information transfer is impossible, because the random process deprives the information of meaning (if one defines meaning as the aspect of information that allows one to act upon it).

A fuller incorporation of the aforementioned frameworks into Kennedy's arguments would probably have resulted in a slightly different model and, more importantly, in slightly different suggestions for future research. With these comments, I would like to point future psi researchers, especially those with a physics background, to further explore the abovementioned theories as a possible explanation for the elusive character of psi.

REFERENCES

BIERMAN, D.J. (1980). Negative reliability: The ignored rule [Abstract]. In W. G. Roll and J. Beloff (Eds.), Research in parapsychology 1980 (pp. 14-15). Metuchen, NJ: Scarecrow Press.

BIERMAN, D.J. (1996). Do psi phenomena suggest radical dualism? In: S. R. Hameroff, A. W. Kaszniak, and A. C. Scott (Eds.) Toward a science of consciousness II (pp. 709-714). MIT Press, Cambridge, MA.

BIERMAN, D.J. (2001). On the nature of anomalous phenomena: Another reality between the world of subjective consciousness and the objective world of physics? In P. van Loocke (Ed.), The physical nature of consciousness (pp. 269-292). New York: Benjamins.

CARPENTER, J. (1991) Prediction of forced-choice ESP performance: III. Three attempts to retrieve coded information using mood reports and a repeated-guessing technique. Journal of Parapsychology, 55, 227-280.

KENNEDY, J. E. (2001). Why is psi so elusive? A review and proposed model. Journal of Parapsychology, 65, 219-245.

MAY, E. (1996). The American Institutes for Research review of the Department of Defense's STAR GATE program: A commentary. Journal of Scientific Exploration, 10, 89-107.

MILLAR, B. (1978). The observational theories: A primer. European Journal of Parapsychology, 2-3, 304-332.

MILLAR, B., & HARTWELL, J. (1979). Dealing with divergence [Abstract]. In W. G. Roll (Ed.), Research in parapsychology 1978 (pp. 91-93). Metuchen, NJ: Scarecrow Press.

UTTS, J. (1991). Replication and meta-analysis in parapsychology. Statistical Science, 6, 363-403.

UTTS, J. (1996). An assessment of the evidence for psychic functioning. Journal of Scientific Exploration, 10, 3-30.

VON LUCADOU, W. (1990, June). Some remarks on the problem of repeatability of experiments dealing with complex systems. Irsee, Germany: NATO Advanced Study Institute, Information Dynamics.

VON LUCADOU, W. (1994). Wigner's friend revitalised? In H. Atmanspacher & G. Dalenoort (Eds.), Inside versus outside (pp. 369-388.) New York: Springer.

WALKER, E. H. (1975). Foundations of paraphysical and parapsychological phenomena. In L. Oteri (Ed.), Quantum physics and parapsychology (pp. 1-53) New York: Parapsychology Foundation.

WALKER, E. H. (1985). On the mathematics of scientific belief systems. In B. Shapin & L. Coly (Eds.), The repeatability problem in parapsychology (pp. 98-143). New York: Parapsychology Foundation.

DICK J. BIERMAN

Chair of Parapsychology

Bestuursgebouw, rm. 271

University of Utrecht

Postbus 80125

3508 TC Utrecht

The Netherlands

d.bierman@bs.uu.nl, bierman@psy.uva.nl

To the Editor:

Dick Bierman is unfortunately correct in gently pointing out that I was not aware of his paper discussing decline effects in parapsychology. In that important paper (Bierman, 2001), he provided compelling evidence from meta-analyses databases that declines in psi effects across studies are pervasive throughout parapsychological research. He also presented several ideas about why this occurs.

Even though I missed his paper by oversight, my paper appears to complement his work because my focus was on the psychological and evolutionary factors that may explain the elusive nature of psi while his work has focused on ideas from physics and systems theory. Of course, there is some overlap. I believe it is very possible that something along the lines that he proposes will turn out to be correct.

Although some may disagree, I tend to think that the psychological factors relating to the elusive nature of psi are more testable and will have to be investigated and discarded as inadequate before the observational theories about multiple potential future observers and the even less testable systems theory approaches will become the favored interpretations. Of course I would not want to discourage Dick and others from trying to dream up ingenious experiments that would directly test these ideas. But, my expectation is that we will have to work through the psychological factors first.

For example, prior to seeing his analysis, I had thought that the declines may be largely a result of unsuccessful replication efforts by non-psi-conducive experimenters, which would mean that relatively stable results may be possible for certain experimenters. But now I'm less sure.

The important question of to what extent the declines are due to decreasing success by psi-conducive experimenters versus unsuccessful replication attempts by non-psi-conducive experimenters needs direct evaluation.

The three factors, namely decline effects, differences among experimenters, and goal-oriented experimenter effects, ultimately must be considered together because each one confounds the other two. With goal-oriented experimenter effects, the tendency for later studies to have larger sample sizes would result in a misleading decline in effect size. With goal-oriented psi, the z score is independent of sample size, which means that the effect size is inversely related to sample size (e.g., effect size is z/[square root of]n. Under the traditional assumptions that the subjects produce the psi effect, the z score increases with the sample size, and the effect size is independent of the sample size. With goal-oriented experimenter effects, the usual meta-analysis approach that focuses on effect size will have a large component of noise because this is not an appropriate outcome measure in this situation. The z score would be a more relevant measure.

It may be possible to sort out these three factors with the existing meta-analyses databases. Bierman has taken the first step by evaluating the main effect for decline effects. The next steps are to evaluate the main effects for experimenter differences and for goal-oriented psi, although available data provide fairly good indications of how these will come out. The final step is to look at interactions and build models.

Another important psychological factor that relates to the elusive nature of psi is to understand the motivations associated with psi. One of the many important aspects of this topic is that observational theories and related ideas assume conflicting motivations among people that suppress psi occurrence with more observers. One of the points in my review was: what evidence do we have for this assumption? This assumption appears to be needed to make the theory fit the data and tends to be accepted without question. However, I have found little independent support for this assumption from either research or personal experience. It is possible that this assumption, and models that depend on it, are on the wrong track.

I believe the motivations relating to psi need to be understood before there is any hope of understanding the occurrence and elusiveness of psi. This is one of the most important and neglected topics in parapsychology. It is also relatively easy to investigate.

As a minor point of clarification, Bierman states that I "incorrectly" suggested that "many or all of the people who are interested" are involved in producing a psi result. He indicates that this cannot be concluded from the experimental data. It may be noted that my paper discussed both sides of this issue and concluded that the relevant data are inconclusive. He and I appear to be in substantial agreement about the overall status of this research.

REFERENCES

BIERMAN, D.J. (2001). On the nature of anomalous phenomena: Another reality between the world of subjective consciousness and the objective work of physics? In P. van Loocke (Ed.), The physical nature of consciousness (pp. 269-292). New York: Benjamins. Also available at: http://rea3140.fmg.uva.nl/PUBS/2001/

J. E. KENNEDY

Boulder, Colorado

72130.1210@compuserve.com