jonathan p. dowling et al- comment on “theoretical model of a purported empirical violation of the...

8/3/2019 Jonathan P. Dowling et al- Comment on Theoretical Model of a Purported Empirical Violation of the Predictions of Quantum Theory

1/12

i

Comm ent on Theoretical Modelof a PurportedEmpiricalViolation of thePredictions of Quantum TheoryJonathan P. Dowling*

Quantum ComputingGrou p, Mail Stop 126-347NASA Jet Propulsion Laboratory, California Institutef Technology4800 Oak Grove*Drive, asa&na, Cali$ornia 91I 0 9 USABerthold G . Englert and Axel SchenzleMax-Planck-Institutjiir QuantenoptikPos fach 1513,0-8 .5748 Garching, Germany

James E. AlcockDepartment of Psychology, Glendon CollegeYork Un iversity, Toron to,M4N 3M6 Canada

Ray HymanDepartment of Psychology, University f Oregon, Eugene, Oregon 97403USA

AbstractWe show that an alternative quantum heory proposed by Stapp directly violates Einstein

causality. This model was constructed to explain a single questionable experiment on paranormalpsychic abilities in humans.

Version2.2,05 NOV 98, submitted to Skeptical Inquierer


2/12

i

Our commentconcerns hearticle by Henry P. Stapp, 'Theoreticalmodel of a purportedempirical violation of the predictions of quantum theory" [11. This paper develops a theoreticalmodel of nonlinear quantum mechanics that is loosely based on work of W einberg [2], and thatturns out to violate Einstein ausality,'as we s h d show below. It is clear that this article has beenspecifically created toexplain the apparentlyanomalousresultsfound n what is essentially asingle, unreproduced, highly questionable experiment designedo establish the physical realityof

paranormalphenomena. In particular, Stapp 's reference 8 is to he elekinesisexperiment ofSchmidt (31, published in the Journal of Parapsychology. (Stapp actually participated nonephase of this experiment,a "sub-experiment" and co-authored n experimental reportwith Schmidt[3]; a fact that is not evident from Stapp's theory paper [l].) The conclusion of Schmidt-thatStapp is trying to model theoretically-is that the experimental test subjects are using their psychicpowers to alter the laws of quantum mechanics and thereby modify the outcomef events that havealready occurredn the past!

In the span of one minute we couldconstructat east 10 theories nvolving 10 separatemodifications o the Schrodingerequation hat wouldproduceoneormore of thefollowingnonphysical effects: oncausality, nonlocality, energy non-conservation, nonunitary evolution ofthe wave function, andparity violation. In the absence of overwhelming experimental support thatthese effects indeed occur, such modifications do not constitute new physical theories bu t rather

2


3/12

mathematical exercises of little value. Should a theory that abandons Einstein causality be true, itwould shatter the very foundations of physics as we know it. Such a modification to our basicphysical laws should be made only if a vast body of repeatable experimental evidence requires usto do so. We shall argue below that'the experiment of Schmidt does not provide this evidence.However, Stapp's model does indeed abandon causality,as we shall see.

The Schmidt articleo which Stapp refers s not a reportof a particular experiment, but ratheri of a meta-analysis of five sub-experiments, including one in which Stapp himself was involved

[3]. Two of the sub-experiments included n the meta-analysis are unpublished and therefore havenot been subjected to peer review.

It is important to understand the basic setup f these sub-experiments in order to comprehendthe ature of the claims beingmade.Theseub-experimentsnvolve ot nlyupposedpsychokinesis (PK)- y which the putative mental influence of the subject can directly, by noknown physicalchannel, nfluence matter - ut a retroactive psychokinetic nfluence thatsupposedly operates backward in time. Although the five sub-experiments differ in some respects,a brief review of the first of them provides a feeling for the naturef the research.

In that first sub-experiment,Schmidtusedacombination of Geigercounterandrandomnumbergenerator ogeneratea istof ix-digit andomnumbers 141. Thesenumberswereemployed in serial order, two at a time. Next, using an electronic noise random generator, Schmidt

3


4/12

produced a series of binary decisions , labeled High and Low. Then he sent the list of seednumbers and the target assignments (High and Low) to colleagues at Syracuse University,along with a program that was based on an algorithm that would use each seed numbero producesome visual output on the computer e e n . (He used three different programs to produce differentvisual displays, but one is enough to communicate the flavor.) The algorithm, starting with theseed number, generated a series of 0s and l s , which were translated into the presentation on the

i monitor of a swinging pendulum, the amplitude f which was determined by the binaryseries.So, the subject sat in front of the computer screen and the pendulum began to swing. If the

target assignment was for High, he would attempt to increase the pendulums amplitude; if it wasfor Low,hewouldattempt oreduce t. After aseries of trialswascompleted, hesubjectscalculated the score for each trial (i.e., a number produced by the algorithm, based on the seednumber, and corresponding to the amplitude of the pendulum). Each pair of seed numbers, then,generated,via healgorithm,acorrespondingpair ofoutcome numbers, one witha argetassignment of High and the other of Low. The difference between the score or the H number andthe score for the L number provided the basic datum. These differences should average to zerosince the seed numbers were randomly generated and the High-Low assignments were randomlygenerated.

4


5/12

The goal was to produce higher numbers for the High target assignments than for the Lowtarget ssignments. A smallbut tatistically ignificantlydifferencerom erowas found,indicating that the pairs of seed numbers had some correlation with the H-L assignments. Thus ,two random processes were correlated. Schmidt and his co-authors were careful o avoid offeringany particular theoretical interpretationof these results. However, in later papers, y Schmidt, andbySchmidt ndStapp (31, one of .the xplanations proffered for such effects is that of

i noncausality, whereby he mental effort of thesubjectretroactively affects the random events.According to this view then, the mental effort of the subjects would have reached back into thepast, to the time when radioactive decay was producing events on a Geiger counter leading to thegeneration of six-digit seed numbers, which when taken in consecutive pairs, subsequently boresome relationship to each other- relationship that wast that point yet to e specified by anotherrandom process, the electronic noise random generator.

Even more amazing, a task that no conscious brain would evex be able to carry out, wassupposedly canied out unconsciously and automatically, across a continent (the subjects were inSyracuse; the number generators were in St. Antonio), and across time- o a particular timeunknown to the subjects. Essentially,all the subjects needed to do was to wish and their wishescame true, to a significant, albeit small, degree. Note that his subexperim ent has never beensuccessfully independently replicated. The otherubexperim ents were of the same genre.

5


6/12

As far a s we can tell from the research report, the empirical methodology do es not by itselfpose a problem- hat is, there are no obvious problems. There are some minor weaknesses. Forexam ple, across his research history, Schmidt has often serveds both subject and experimenter ina given study, and indeed, this as the case in some of the research under consideration here.Thisis not considered appropriate in orthodox psychological research, but neither does it automaticallyinvalidate hisdata. However, given that there has beeno replication in an independent laboratory,

i it is impossible to pronounce that hemethodologywasbeyond eproachbasedonlyon heresearch report. Flaws and biases in psi research typically are not recognized as such by heresearcher and therefore there s usually no evidence of them in the researcheport.

We are more concerned with regard to the apparent arbitrariness of the statistical ana lysis.Schmidthasused a particularmethod of meta-analysis hat,whilenot ncorrect, does raisequestions as to why the particular method was chosen, and whether or not other methods wouldyield similar esults. The standard methodof meta-analysis is to begin by calculating an effect sizeand then o connect effect sizes to probabilities. Why this was not dones not clear.

Wenote hat Schmidtemploys a ratherunorthodoxwayofbreakingup hedata ntoindividual units within a sub-experimenL It is not clear jus t why this was done, or what theimplications of this procedure are or the subsequent data analyses.n order to assess the propriety

6


7/12

of the data analysis in general and the meta-analysis in particular, it would be necessary to haveaccess to the raw data.

Historically, the results of successfulparapsychological indingshavenotheldup -especially across independent laborakries and across generations. In orthodox fields of scientificinquiry, the standard practice is to withhold judgment until - a novel finding- specially one thatchallenges prevailing theories - s independently replicated. The criterion of proof seem even

i more mportantwhenconsideringaclaim hatchallengesoneof he undamentalpremisesofmodern science, namely Einstein causality.

Suffice it to say that no independent laboratory has replicated any of these sub-experiments.Given hat, as Schmidtadmits,most of he ndividual subexperiments themselves failed toproduce statistically significant results, the Schmidt data does not in any way confirmhe existenceof apsychokineticeffect.Thus, here is nobasis orarguing hataviolationofquantummechanical theory has been demonstrated.

In the history of Schmidts PK research, an acausal explanation of such experimental resultshas been offered. In particular, in the Schmidt and Stapp experimental paper, we find3]:

The esults of [psycho-kinetic]experimentswithpre-recorded andomeventsappear most interesting and most puzzling because the subjects mental effort ismade ongafter he andomevents o be affectedhaveoccurred.One entativeviewpoint [5,6] is that the subjects mental effort could act backwards to the timewhen the random events were generated and recorded. This would imply a non-

7


8/12

causal mechanism in the sense that the effect (the biasing of the random events)occurs before the cause (the mental effort).t might be this element of noncausalitythatmakespsychicphenomena so intuitively mplausibleandat odds with heknown principlesof physics.

It appears that in additiono this hypothesis, a second was eventually given-perhaps becausesuch a blatantly acausal mechanism as outlined above is considered utterly implausible to mostphysicists as it is inconsistent withEinsteinian elativity.Hence, as a subs titute, a econd

i explanation was offered [7], viz . that the subjects bias the quantum-mechanical state reduction atthe ime when heyexercise heirmentalpowers. In thisway, it is hoped,anyviolations ofcausality are avoided, simply because the subjects are not supposed to influence the radioactivedecay months ago but nly the state reduction that allegedly happens at thetime of each of the PKsub-experiments. It is apparently on this second hypothesis that Stapp models in his paper11.

There is much to be said about the fundamental misunderstanding of the quantum mechanicalnotion o f state reduction nherent n this hypothesis. In particular, he ar-fetched implicitassumption is simply preposterous, namely, that one has a coherent entangled wave function forthe complete system o f : adioactive sample, Geiger counters, monitoring PC, floppy disk drive,floppy disk, printer,print-out,UnitedStates Postal Service, plus anything else hat has beeninteracting with any one of the componentsn the time between the clicksof the Geiger counter andtheactual PK experiment. Cons ider -even in hebestEPR-typeexperiments it is extremely


9/12

difficult to get microscopic, entangled quantum states of matter that extend just across a lab benchfor a few microseconds.

We shall finish our comment by pointing out that Stapps theoretical model of Sch midtsbizarre expcriment is intrinsically seif-contradicting. Stapps model is technically based upon thenonlinear extension of quantum mechanics studiedby Weinberg [2,8], but really the nonlinearitiesof Weinbergstype are not essential. One can revert to ainear theory, as does Stapp [l].What is

i essential is a non-Hermitian piece of the Hamiltonian which givesrise to a nonunitary evolution.At first sight, the avoidance of Weinbergian nonlinearities takes care of Gis ins objection [9]

who observed that nonlinearities of any kind in the Schrodinger equation always enable one toconstruct a mechanism or sending messages faster than the speedof light. And as soon as this ispossible, acausal actions backwards in ime are possible as well [lo]. Incidentally, the note at theendof eference 8 clearly ndicates hatWeinberghimselfconsiderssucha state of affairsunacceptable.

A second look however, reveals that Stapps proposalf a linear, yet non-Hermitian evolutionis just as acausal as is the nonlinear, Hermitian one. The reason is the necessary normalization ofthestatevector to unit ength,needed or hecorrectcomputationofexpectationvalues orprobabilities. The complete process of (i) a linear, non-Hermitian evolution, followed by (ii) thenormalization of the state vector, is a nonlinear mapping (not necessarily f the kind considered by

9


10/12

Weinberg)of he nitialstatevectoron he inal one.Consequently,Gisin'sargumentsaboutnonlinear quantal evolutions apply, and Stapp's model is recognized to be unavoidably acausal.Therefore, Stapp was not successful in his attempt at constructing a worlung modelor Schmidt'shypothesis of an explicitly causal mechanism hat biases he state reduction. In sho rt, Stapp'smodel is just as inconsistent with Einsteinian relativity as was Schmidt's original and untenablehypothesis [3,5,6].

Inconclusion, hemodelproposed byStapp smanifestly nconsistentwith heknownphysical law of Einstein causality. As such, it should be taken seriously only if t explains anoverwhelmingbodyof ncontrovertibleexperimentalevidence ndicating hatsuchviolation ofphysical law actually occurs. As we have pointed out above, the single experiment that seems toindicate hisviolation, sextraordinarily"underwhelming",andweconjecture hat twillbeunrepeatable by any independent experimental group.

A C KN OW LED GMEN TSThe authors of this work would like to acknowledge useful and interesting discussions with

Benjamin Bederson, Bernd Crasemann, Daniel Greenberger, Lorenzo Narducci, Vic Stenger, andAnton Zeilinger. Oneof us (JPD) would also like to recognize that a portionof research describedin this paper was carried ut by the Jet Propulsion Laboratory, California Institute of Technology,under a contract with the National Aeronautics and Space Administration.

1 0


11/12

1.

2.3 .

4.

5 .6 .

7 .

8.

REFERENCESH. Stapp, TheoreticalModel of aPurportedEmpiricalViolationof hePredictionsofQuantumTheory," Phys.Rev. A 50, 18 (1994); J . P. Dowling,(Letters) PhysicsToday48, 78(July 1995).S . Weinberg, "Precision Tests of Quantum Mechanics," Phys. Rev. Lett. 6 2 , 485 (1989).H. Schmidt and H. Stapp, "Study of Psycho-K tnesis With Pre-Recorded Random EventsandheEffects f Pre-Observation," J . Parapsychol. 57, 331 (1993); H. Schmidt,"Observation of a sycho-KineticEffectUnderHighlyControlledConditions," J .Parapsychol. 57, 351 (1993).H. Schmidt, R. L. Morris, and L. Rudolph, "Channeling Evidence for a Psycho-KineticEffect to Independent O bservers," J. Parapsychol. 50, 1 (1986).H. Schmidt, "Can Effect Precede Its Cause?"Found. Phys. 8, 463 (1978).E. H. Walker, "The Quantum Theoryof Psi Phenomena," Psychoenergetic Systems 3, 259(1979).H. Schmidt, "Collapse of the State Vector and Psycho-Ktnetic Effect," Found. Phys. 1 2 ,565 (1982).S . Weinberg, "Testing Quantum Mechanics, "Ann. Phys. (N Y) 194, 336 (1989).

1 1


12/12

jonathan p. dowling et al- comment on “theoretical model of a purported empirical violation of the...

Documents