sign in sign up
<<
Home , Psychometry (paranormal), The Mentalist

--'N~A~Tu~R'-'-E~v~o~L--'. ""'32"'0_1"'3--'M""""A'-'-R'-'-c"""'H-'--'-1"'"'9s"'6------(QMMENTARY------ll_9 Investigating the

from David F Marks

Parascience has so far failed to produce a single repeatable finding and, until it does, will continue to be viewed as an incoherent collection of belief systems steeped in fantasy, illusion and error.

FEW fields of inquiry capture the attention ignore such developments is, to say the were claimed to be paranormal but which of the public as much as the paranormal. least, uncharitable. can now be explained from within Newspapers, books, films and television The Committee for the Scientific orthodox science include: have all cashed in and promoted it. Yet Investigation of Claims of the Paranormal (1) , the photo­ after millennia of experience and more (CSICOP) was established in 1976 with graphic recording of coronal discharges than a century of controlled investigation, the aim of increasing the quality of scien­ around living or non-living objects pro­ since the founding in 1882 of the Society tific investigations into the paranormal by duced by high-voltage, (20-100 kV), high­ of Psychical Research, the paranormal constructive criticism and the exposure of frequency (75 kHz-3 MHz) electrical 13 14 remains as controversial as ever. While invalid or fraudulent claims. Over this 10- fields • . Variations in the images of the credence in extrasensory (ESP) year period, an inordinate amount of corona can be explained in terms of nor­ and is widespread, fraud, error and incompetence in paranor­ mal physical factors such as moisture, has failed to produce a mal investigations has been brought to pressure or distance, all of which influence 11 single repeatable demonstration. In the light'- • But pseudo-sciences are remark­ circuit resistance. face of such a dearth of hard evidence, ably stable and tradition-bound; their (2) Fire-walking, if conducted briskly how can such widespread belief in the presence on the edges of science can be on hot materials of low thermal capacity 12 paranormal be accounted for? expected indefinitely . and poor thermal conductivity, does not 15 16 The importance of rigorous analysis of Areas of experimental psychology can produce burns • . The Leidenfrost effect the evidence for parascientific claims can­ shed light on the paranormal, especially created by an insulating layer of water or not be underestimated. The establishment the study of consciousness and cognition. sweat may also reduce transfer to of ESP could conceivably require a para­ Such investigation indicates that the many the surface of the body. digm shift (in Kuhn's sense) of the most anomalies of putatively paranormal (3) is based on sensory cues, fundamental kind and our concepts of experience are an inevitable consequence expectancy effects and probability. Con­ mind-brain relationships and conscious­ of normal selective and constructive pro­ trolled trials fail to produce above-chance 8 11 19 ness would need radical alteration. Our cesses in perception, memory and results • - • whole approach to psychology as an imagery. I summarize here what seem to (4) surgery, thought photogra­ empirical science, based as it is on the be the common assumptions on which phy and metal bending all involve sleight­ time-honoured assumption that percep­ claims of the paranormal are based. of-hand and can be duplicated by skilled 8 20 22 tion can result only from sensory activity, magicians • - . The first differs from the would be brought into question. Theoretical assumptions others in respect of the associated false The conventional response of many Paranormal phenomena are negatively hopes and financial loss, but all three are scientists to the paranormal is to ignore defined. A phenomenon is defined as para­ fraudulent. 2 1 22 the evidence on a priori grounds, believing normal only if it contravenes some funda­ (5) 'Gellerized' watches • , thought to it to be of basically poor quality. This mental and well-founded assumption of be broken, are purportedly repaired by attitude is allied to the Humean stance science. Hence, to establish an effect as illusionist by 'psychic con­ that a lie is more probable than a . paranormal, all possible ' normal' explana­ centration'. In about 50% of cases, simply Although this scepticism is certainly jus­ tions must be shown to be invalid. Any holding the watch in a clenched first and tified, it could be argued that such a paranormal claim thus also remains pro­ shaking it provides a sufficient stimulus to 23 blanket response is counterproductive. visional; a normal explanation, not pre­ free the mechanism . First, it is hardly scientific to reject a claim viously thought of, may be discovered at (6) , graphology, tea-leaf and purely because of its a priori improbabil­ some future time. tarot card readings, the I Ching and other ity. Second, a division is created between Mysteriousness per se is a necessary but forms of are all types of 'cold 24 aligned groups of committed 'believers' insufficient condition for adjudging an reading' or 'sleight-of-mouth' . They and 'sceptics' and the resulting adversarial event as paranormal. There will always be depend upon ambiguous, wish-fulfilling positions inhibit proper discourse and the limits to knowledge, so that new and general advice, the use of prior or possibility of an account which satisfies phenomena that initially appear presented information and cues obtained all parties. Third, it leaves the field open anomalous will be given a natural expla­ by verbal 'fishing' . A strong feeling of per­ for undisciplined exploitation, which is nation following systematically controlled sonal validation often accompanies such irresponsible; there are many examples of observations. Bona fide paranormal readings. Various forms of '' financial loss, suffering and even death effects, on the other hand, are supposed and 'psychometry' as practised by D. Col­ resulting from fraudulent paranormal to contravene established assumptions as lins and D. Stokes are also examples of claims (for example, the Jonestown though from another order of existence . massacre, , the Transcen­ and not simply for lack of explanation. In some cases, field observations can be dental programme, 'Contranormal' would be a more precise checked under laboratory conditions and firewalking, scientology and other pseudo­ technical term. the sensori-motor features of the original scientific cults). For scientists passively to Examples of effects which until recently performance reproduced using a delayed _12_o______CQMMENTARY ------N_A_T_uR_E_v_o_L_. 3_2_0_1_3 _M_A_R_C_H_t_9_86 control group of non-psychic subjects; for example, Geller's watch-starting pro­ a cedure and ability to draw the contents of sealed and a ppa rently opaque envelopes were matched by that of non-psychic con­ trols23 (Fig. 1). Clearly, the tendency to judge a mysterious event as paranormal in the absence of controlled observations can be quite misleading. b The most dramatic evidence for the paranormal has been based on either fraud or methodological error. Apparent frauds that have been uncovered include Univer­ sity of London mathematician S. G. Soal's 25 27 manipulation of his recording sheets - , c University of Utrecht Professor Tanhaeff's evidence on Croiset, the Dutch 'psychic' detective28 , and the description by C. Cas­ taneda (University of California at Los Angeles) of the paranormal teachings of Don Juan29 . C. E. M . Hansel 3 has provided a valuable review of the history of trickery, Fig. I Drawings presented to Geller and to non-psychic controls. The targets were inside fraud and error in parapsychology. envelopes, and Geller and the controls were allowed to see the sealed envelopes. The first and However, outright fraud is not the only se cond co lumns show targets as they appeared in unfolded and folded states. Geller's final attempts vehicle in which the paranormal cause can at reproducing the drawings, which were folded and sealed inside envelopes, are shown in column travel, and it is a serious mistake to assume 3. Columns 4 and 5 show the best results obtained from the co ntrol subjects after only 5 and I 0 s it is a necessary part of any paranormal of close visual inspection. The drawings were presented inside the same envelopes under similar investigation. lighting conditions. Geller, who claimed to use ESP, had taken 11.5, 85 and 18 min for a, b and There are no theories to account for para­ c, respecti ve ly, but he was being observed, albeit discontinuously (see ref. 23). normal effects or their properties. There are some undesirable implications of this aspect. First, investigators are unable to they themselves believe in psi, the so­ somehow discriminates its own mental conduct properly controlled experiments called 'sheep-goat' effect32 . Hence, psi can experiences from that of all others38 . But on the properties of psi phenomena be held to be pre>ent whatever the results, even putting that issue to one side, it is because they have no idea what the rel­ unless the belief of the investigators is curious how seldom the anti-materialist evant variables are. In particular, there is itself independently controlled. assumption has been properly explained. no procedure by which psi can be deliber­ In an extreme version of the 'sheep­ D. E. Cooper indicated one way in which ately switched on or off, and so there is goat' hypothesis, some investigators have the a nti-materialist argument can be con­ no possibility of examinng the effects of even proposed that the participants in an structed as a reductio ad absurdum, but he 39 psi on other variables. All that can be done experiment need not be restricted to the found this to be incoherent . In fact, it is to establish whether a given perform­ individuals in the laboratory but could seems doubtful that materialism and ESP ance in some particula r set of circum­ also include all of the readers of the jour­ would be incompatible, should real stances differs from a baseline; if so, psi nal which subsequently publishes the evidence of ESP ever be found. As M. is assumed to be the cause. results33 ! In this case the proportion of Scriven has pointed out, materialism can Oddly, neither the subject nor the sceptics and believers among the eventual always be enlarged to absorb newly sub­ experimenter can state which of the suc­ readership would determine the experi­ stantiated phenomena, "since the very act cessful trials in a psi experiment result mental outcome through backward cau­ of substantiation demonstrates that the purely from chance guessing and which sality. Similar in kind is the 'shyness phenomena are indeed part of the material are generated by psi, so that there is no effect', the tendency of metal not to bend world, and hence that a current version of 34 40 basis for distinguishing between a high psychically while it is being observed . materialism must embrace them ." score in a psi experiment and a lucky run Evidence of the paranormal is held to be in a game of chance. Also, the persistence incompatible with materialism. Inves­ Methodological problems of psi investigators in the face of variable tigators throughout history have been con­ The failure of pa ra normal investigators to but mainly negative results could have a vinced that evidence of the paranormal produce a single repeatable effect despite similar motivational basis to that of addic­ proves that materialism must be wrong. I 00 years of published research is a serious ted gamblers; both show high resistance This was assumed by the Society of Psy­ matter. The hoped-for results have been to extinction following variable ratio chical Resea rch, one of whose early presi­ described in thousands of reports, but not 30 schedules of reinforcement . dents, Sir Willia m Ba rrett, spoke of one can be repeated in a properly control­ A more fundamental problem with the parapsychology 'as the most valuable led replication. Yet in addition to the huge 35 36 paranormal's atheoretical status is that of handmaid to religion' . J. B. Rhine , the literature of unrepeatable findings, there untesta bility. Failure to observe a par­ founder of the Parapsychology Associ­ is an inestimable number of unpublished ticular effect can be readil y attributed to ation a nd C. Tart37, a former president, and presumably negative results. a host of ad hoc, hypothetical factors. Vi vi d have both reiterated the religio-spiritual The most systematically investigated imagination is no substitute for testability, motive for pursuing psi research. area is undoubtedly parapsychology. The however, and if ad hoc hypotheses are not An immaterial '' has passed out of field is professionally organized, with its •ndcp~n

a ------.. Freud's interpretation of dreams, preju­ .. -- dice, faith-healing, the placebo effect, bone pointing and the ''. Beliefs of all kinds tend to be self-perpetuating. Coincidences. Psi phenomena consist of an experience, image or thought matched by some other similar experience, image or u,5l... --r _ ,p-- thought. Collections of such coincidences 62 have been published by A. Koestler , L. ~~~-~=-';;;;i', .~-~ Rhine,63 and others based on the assump­ --'"'--==?"------tion that odd-matches of events cannot occur purely by chance. Probability theory shows that an event which is improbable over a short run can become highly probable over the long run. If five coins are tossed all at once on a single occasion the probability of obtain­ ing five heads is T 5 or approximately 0.03. If the coin tossing is repeated 100 times the probability of five heads somewhere in the series is approximately 0.96. The principle of the long run is easy to grasp in simple situations but much less visible in the more chaotic world of spon­ taneous human experience. Calculation Fig. 2 a, A remote viewer's drawing of the target site shown in b, an A-frame house and steps, shows how easily Koestler could obtain purportedly using ESP. When taken to the target area, the subject was delighted when he discovered his 40-plus odd-match anecdotes. Assum­ the view from above the house, halfway up the steps (c). When an independent judge visited the ing that in an ordinary day a person can 10 target he was convinced that the correctly matching drawing was d. This had actually been recall JOO distinct events, there are °C2 produced for another target site, a railway station (see ref. 4). or 4,950 pairs of events per day. Odd­ matches can be remembered for years, per­ haps I 0 yr or 3,650 days. If Koestler knew images can be easily misinterpreted in mentalists's tricks cons1stmg of: (I) cor­ 1,000 people, he could draw upon a total 57 terms of pre-existing beliefs . rectly naming a colour written out of sight; pool of 4,950 x 3,650 x 1,000, or more than Expectancy or mental set provides the (2) correctly transmitting a colour name 18 x 109 pairs of events. That Koestler framework within which we organize new to a volunteer who, like me, had not pre­ obtained 40 striking odd-matches seems experience. Human cognition is not a viously seen it; (3) helping a volunteer hardly surprising. simple copying process but entails a con­ correctly to read messages sealed inside Koestler's (see ref. 4) is certainly structive striving or 'effort after meaning'. envelopes or to appear to transmit not unique to him, although he was one What we experience is often more a messages to me; (4) producing bent keys of a small group of analysts who wanted confirmation of belief than a matter of which I had not previously touched; and to make a scientific revolution out of it. plain fact. Beliefs are not automatically (5) moving or stopping the hands of a The fallacy is widespread and several updated by the best evidence available, watch in a mysterious manner. biases contribute to it. First, we notice and but have an active life of their own and Although at no time did I claim to be remember odd-matches. Second, we do fight tenaciously for their own survival. psychic, 90% of the class stated that I had not notice non-matches. This triggers the They tell us what to read, what to listen demonstrated psychic ability. When the short-run illusion that makes the odd­ to, who to trust and how to rationalize results from subjects who had previously match seem improbable. Third, we are 4 5 57 contray information · • • been classified as 'believers' and 'sceptics' normally poor estimators of probabilities, Selective exposure protects beliefs from were analysed separately, 79% of believers especially for combinations of events. more dramatic forms of contradiction. thought at least three of the five effects Unseen causes. Another class of psychic­ When U. Geller visited the were psychic compared with only 43 % of looking experiences is generated by invis­ city of Dunedin in New Zealand there sceptics (P<0.001). ible chains of cause and effect which bias were seven different opportunities to Naturally, we often encounter informa­ the probabilities away from chance levels. obtain information abut his alleged psy­ tion that is unexpected or ambiguous. In Failure to randomize target stimuli chic abilities: four media interviews, two such instances, there is a second line of properly in ESP experiments is a good newspaper stories and one stage perform­ defence: the data can be selectively per­ example of this. Thus, Tart reported a ance. Of 17 subjects who, before Geller's ceived or even misperceived so that they successful ESP experiment in which his visit, were already 'believers', 15 selected still appear to support our beliefs by 'sub­ subjects learned to score above chance in 4 three or more of the available exposures. jective validation· . One illustration of this guessing which of 10 digits was displayed Of 20 'non-believers', only 10 selected as powerful cognitive defence in the context by an apparatus in another room following 2 many as three exposures (x ( 1) = 6.13; of ESP research is the strong conviction the presentation of feedback),. The ran­ P <0.02). that one has successfully viewed a com­ dom number generator mistakenly A further problem is that when we are plex target site by ESP in a remote-viewing avoided using the same digit twice in suc­ exposed to relevant information, our experiment even when one is completely cession, a bias which is matched by the opinion revisions are often less than wrong ( Fig. 2). pervasive 'gambler's fallacy'. When Tart optimal, and we act like conservative There are many now-classic exampies removed this bias, the 'ESP' also disap­ 5 59 Bayesians ~ , with a . In of subjective validation: the prophecies of peared65. 60 a recent ' ESP' demonstration to a class of the Delphic Oracle and , Another unseen factor, used by illusion­ 61 226 psychology students, presented as an the discove ry of N-rays , phlogiston, Vul­ ists, is the 'population stereotype'. The exercise in observation, I performed five can, the canals on Mars, fl ying saucers, performer 'sends a message to the NATURE VOL. 320 13 MARC H 1986 COMMENTARY 123

>60 a 240 h between believer and sceptic; one could well be arguing about the existence of God. Belief in the paranormal is metaphy­

O> 140 sical and therefore not subject to the con­ "'c: 200 "'c: .c"' straints of empirically based science. u .9 Parascience has all the qualities of a £ .~ a.'" magical system while wearing the mantle "'() E 160 c: '20 of science. Until any significant dis­ ;;; 0 ·;;;"' coveries are made, science can justifiably ci ~ c: ignore it, but it is important to say why: "'C1l 120 parascience is a pseudo-scientific system 100 ,__0"' "'Q; of untestable beliefs steeped in illusion, > «: error and fraud.

BO I thank Jerry Andrus, Bob Audley, Ray 80 Hyman, A. R. Jonckheere, Peter McKel­ Sub Super Beliefs Sub Super Beliefs lar, J. Randi, Christopher Scott, Jean Constructs Constructs Waugh and many colleagues in CSICOP Fig. 3 a, The average resistance to change scores for 'believers' (B) and 'sceptics' (S) for for useful discussions and information. subordinate (Sub), superordinate (Super) and belief constructs. Believers showed a significantly The late Richard Kammann contributed higher resistance to change for superordinate constructs than sceptics. b, The total number of substantially in the earlier stages of this implications for the same sets of constructs. The interaction between group and construct-type research. was highl y significant ( F (2, 42) = 9.14, P < 0.005 ). Believers saw significantly more implications in changing positions on superordinate constructs than did the sceptics, while sceptics saw David F. Marks is at the Department of significantly more implications in changing their beliefs than believers ( P < 0.05 for both com­ Psychology, University of Otago, PO Box 56, parisons). Dunedin, New Zealand.

I. Kunz, P. 3. 14-32 ( 1978). 2. Diaconi s. P. Science 201, 131 -136 ( 1978). audience, saying "I am thinking of a num­ system of constructs with which people, 3. Hansel. C. E. M. ESP and Parapsychology: A Critical objects and events are compared, con­ Reeva luation ( Prometheus, Buffalo, 1980). ber between 1 and 50, both digits are odd, 4. Marks. D . & K ammann, R. The Psychology of the Psychic and different". Controlled experiments trasted and predicted. Core constructs {Prometheus. Butfalo, 1980). show that the most common response for have relatively superordinate positions 5. Alcock. J. E. Parapsyc·hology: Science of Magic'? (Pergamon, Oxford, 198 11. the 1-50 problem is 37, which accounts and a large range of convenience while 6. Frazier. K. {ed .). Paranormal Borderlands of Science ( Prom­ for 30-35% of all responses, and the peripheral constructs are relatively subor­ etheus, Buffalo, 1981 ). 7. Gardner. M. Science Good Bad and Bogus {Prometheus, second most common response is 35 (20- dinate and more easily altered. Waugh Buffalo. 198 1). 4 25% ) • ff the performer always says he <.ompared the personal construct system~ 8. Randi. J. Flim- Flam ! Psvchics, ESP, Unicorns, an,; 01her had been thinking of 35 and then changed of sceptics and believers in the paranormal Delusions ( Prometheu~. Buffalo, 1982). 9. Kurtz. P. Skep1ical Inquirer 8. 239-246 ( 1984). his mind to 37, at least 50% of the using a belief questionnaire. Ten subor­ 10. Frazier. K. (ed.) Science Confronts 1he Paranormal {Prom­ audience will be thinking of the ·correct' dinate and ten superordinate67 constructs etheus. Butlalo, 1985). 11. Kurtz, P. (ed.) A Skep11cs Hat1dbook of Parapsychology number. were generated using standard procedures (Prometheus, Bulfalo. 1985). Human beings never behave randomly. and each subject's constructs were tested 12. Bunge, M. Skeptu.:al lt1q11irer 9. 36-46 (1984). t3. Tiller, W. A. NewSciem. 62.160- 163 ( t974). Our experiences contain many culturally for their relative resistance to change and 14. Pehek. J. 0 .. Kyler, H. J. & Faust. D. L. Science 19.a. 263-270 shared elements such that particular items the number of implications entailed by (19761. are associated with particular verbal con­ changing the subject's preferred pole on 15. Houdini Miracle M o'1gers and TI1 eir M ethods ( Pro metheus, Buffa lo, 198 1 ). texts. This causes associative networks to the 20 constructs and 10 paranormal 16. Leikind, B. J. & McCarthy, W . J. Skeprical In quirer IO, be set up and a tendency towards non­ beliefs (Fig. 3). 23 -34 ( t985). 17. Vogt , E. H. & Hyman. R. Wa1 erwitd1ing USA 2nd edn random, stereotypical responses even Believers' core constructs were sig­ (Chi cago University Press. 1979). when there is freedom to choose. nificantly more resistant to change and 18. Ra ndi. J. Skeptical lnqwrer 8. 329-33) ( 1984 ). 19. Martin. M. Sk.ep1ical Inquirer 8. IJ8- l40 ( 1983). Other unnoticed causes of putatively there was a parallel difference in the num­ 20. Ra ndi. J. 111e Magic of Uri Geller ( Ballantine. New York, psychic effects include subliminal and ber of implications resulting from changes 19751. 66 at the superordinate level. Compared with 21. Fuller. U. Confessions ofa Psycl1ic ( Karl Fulves. Bo x ·BJ, non-verbal sensory cues which may Teaneck, New Jersey, 1975). lead to common thought patterns in sceptics, believers seem to possess much 22. Fuller. U. Further Confessions of a Psychic ( Karl Fu Ives ~ different people, presenting the illusion of tighter construct systems in which any New Jersey, 1980). 23. Marks, D. & Kammann, R. Zetet1c 1(2). 9- 17 ( 1977 ). telepathy. change at the core level implies a sig­ 24. Hym jn, R. Zerwc t (21. 18-37 ( 19771. The 'will to believe'. What factors nificantly greater upheaval or threat. 25. Saal, S. G. & Goldney, K. M. Prue Soc. psychic:al Res. 47, lt-t50 ( 1943). differentiate believer from sceptic? Waugh also found that believers had sig­ 26. Scan. C. & Haskell, P. Na tu" 245, 52-54 ( 1974 ). Psychologists down the ages have puzzled nificantly higher neuroticism scores than 27. M<1rwick. B. Proc Sue p .. yd1ical Re .~ . 56. 250- 28 1 { 1978). sceptics (see also ref. 68). These data are 28. Hoeben:i . P. H. Skep1ical lnqurrer 6, 32-40 ( 19 ~ I ). over the question of what motivates 29. De Mill e, R. Castandeda's Journey 2nd ed (Capra, Santa different world-views and the so-called congruent with those reported by Zusne Barbara. 19 78). 57 will to believe. Research conducted by J. and Jones who found that believers are JO. Skinner, ll. F. Science and Human Beh avior ( ~1a cm ill an, New York, 1953). Waugh used Kelly's personal construct less flexible than sceptics when confronted JI. Bunge, M. Method, Mode/a,1d M atter ( Reidel, Oordrecht, 6 7 theory. In this framework , people vary with disconfirming evidence. Content 19731. 32. Schmeidler, G. R. & McConnell . R. A. ESP and Pennnaliry in the quality and extent of their investiga­ analyses of believers' construct systems Parrems (Yale University Pre-.s. 19581. tory procedures so that, while some may indicate the presence of spiritual, non­ 3J. Coll ins. H. M. & Pinch. T. J . Frames oj Meaning: 111e Social be working to establish an ordered and materialist constructs at superordinate CmH/fllClum uf Extraurduiury Sden('e ( Rou1ledge & Kcgan Paul. London. 1982). meaningful world which is not highly pre­ level. Such core constructs are not easily J4. Taylor, J. S11rermuuk An ln11wry into the P1.uanormal dictable or readily explained, others may shaken because they are closed off from ( Mai.:mi llan, London. 1975 ). 35. Barratt. W. Prue:. Sue. psrchu:a/ H.1•s. J.i. 275-297 ( 192-l). be content that they already have all the empirical considerations and appear to be 36. Rhine. J. B. Tii e H. ea1·h oj ,\fwcJ. 209-214 t Sl oJne. New necessary explanatory constructs. impermeable to rational persuasion. York, 19471. Hence the feeling of futility experienced J7. Tart, C. T. P-.i: Sl'ierrriticStud1esof 1he P')Thic Realm. vii-viii In Kell y's theory, each individual deals ( Ou1ton. New York. 19 77). wi th the world in terms of a hierarchial in tryi ng to hold rational discussion JM. Fl ew, A. in Snt•nce, Pw11dn- Snencea11d Sonef\· l ed ~ Hanen. _12_4 ______REVIEWARTICLE------N_A_T u_R_E_ vo_ L_._3_20_ 1J_M_ A_ R_c_H_ 19_86

M . P. . Osler, M . J. & Weya nt, R. G.) 55-75 ( Wilfri d 49. T a rt , C. T., Pu thoff, H. E. & T arg. R. N ature 28_., 19 1 ( 1980). Shortco mings of Social J11dgmenr ( Pre ntice- Ha ll , Engle­ Laurier Universitv Press. Waterl oo. 1980). 50. Marks, D. F. Skeptical Inquirer 6. 18-29 ( 1982). wood Cliffs, 1980). 39. Cooper. D. E. in Pliilo-;oph\' a nd Ps\·chical Research (ed. 51. O"Keefe. D. SlfJ/en L1ghtnu1g ( .,on. Oxford. 1982). 60. Hoebens. P. H. Skepr1ca/ Inquirer I, 38-45 ( 1982). Thakur, S. C.) 59-80 (Allen & Unwin. London. 1976). 52. Wilson, S. C. & Barber. T. X. in Image"': C urrent TI1eory, 6 1. Kl ass, P. J . Zetet1c 2. 57-61 ( 1977). 40. Scri ven, M. in Philosophy and Psychica l Research (ed. Research and Appl1ca 11on (ed. Sheikh, A. A. ) 340-387 62. Koestle r, A . Tiie R oots of Coincidence ( Hutchinson. Lo n­ Tha kur, S. C.) 18 1- 194 (All e n & Unwin. London, 1976). (Wiley. New York. 1983). don, 1972). 41. Wolman, B. J. (ed.) Handhook of Parapsychology {Va n 53. Marks. D. & Mc Kellar. P. J. meri taf imagery 6. 1- 124 ( 1982). 63. Rh ine. L. E. J. Pa rap.<.ichol. 15. 164- 190 ( 195 1). Nostrand. New York , 1977). 54. Gurne). E. & M yers. F. W. H. Proc. Soc. psychical R e~. 5, 64. Tart , C. T. Learn ing 10 Use Extrasensory Percep1iu11 ( Univer­ 42 . Beloff, J. Zetetic Scholar 6. 90-94 ( 1980). 403-485 ( 1889). sity of C hicago Press. 1976 ). 43. Morri s. R. L. J. A m. Soc. psy chical Res. 14, 425- 443 ( 1980). 55. Sidgwick, H. Proc. Soc. psychical R es. 10. 25-422 ( 1894 ). 65. Ta rt , C. T. . Palmer. J . & Reding1o n. D . J . J. Am. Soc. 44. Ake rs, C. in Adrnnces in Parapsychological Research Vo l. 56. Finucane. R. C. Appea rance~ of the Dead ( Prometheus, psychical Res. 73. 151-1 65 ( 1978). 4 (ed. Krippner. S.J 11 2- 164 (McFarl and, Jefferson. Buff3lo. 1985). 66. Dixon, N. F. Preconscwus Processing I Wiley. C hi cheste r. No rth Caro li na, 1 98 ~ ). 57. Zusne. L & Jones. W. H. A nomalistic Psyclzology ( Erlbaum. 1981). 45. Hyman, R. J. Pa rapsyclwl. 49, 3-49 ( 1985). Hi llsda le. New Jersey. 1982). 67. Fransell a, F. & Bannister, D. A Mariual fo r Repertory Grid 46. Targ. R. & Pu th off. H. E. Nature 252. 602-607 ( 1974). 58. Edv.ards. W. in Formal Representation of Human Judgem enr Tecllri ique (Academic, Londo n, 1977). 47. Ma rk s, D: F. & Ka mmann, R. Nature 274, 680-681 ( 1978). (ed. Kleinmuntz. B.) 17-52 (Wiley, New York . 1968). 68. Wind holz. G. & Diamant. L. Bull. psychon. Sci. 3, 125-126 48. Hyman. R. Skepr1cal Inquirer 9. 125- 145 ( 1984-5). 59. Ni,.beu. R. & Ross. L. Human Inference: S trategies and ( 1974). ------REVIEW ARTICLE------Heavy-electron metals Z. Fisk, H. R. Ott*, T. M. Rice* & J. L. Smith Center for Materials Science, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA

A new class of metals has been found in which the electrons have effective masses orders of magnitude larger than the free-electron mass. Some of these metals are superconducting at low temperatures. This superconductivity seems to be unconventional, with an underlying mechanism different from that in all other known superconductors.

AMONG the main developments in physics since the 1930s have state properties are dominated by nearly localized electrons. been the discovery and exploration of new ground states of Various features of this superconducting state are unusual, .:ondensed matter. Each of these discoveries opened a new leading theorists to speculate that not only is the mechanism chapter in the physics of condensed matter; recent examples driving this superconducting transition unconventional, but also are spin- and charge-density waves in metals, the supertluid that the configuration of the superconducting state is different state of liquid 3 He and the quantized Hall effect. Very recently from that of an ordinary superconductor. Instead of an interac­ an exciting new class of metallic materials has been discovered tion between electrons that is mediated by phonons (lattice­ with remarkable properties and showing signs of further new vibrational qua nta), which leads to an essentially isotropic gap ground states. (A general compilation of data, with references in the spectrum of electronic excitations, it is envisaged that a current as of mid-1984, can be found in ref. 1.) Coulomb interaction between heavy electrons induces a super­ The terms 'heavy-electron metal' and 'heavy-fermion system' conducting state in which the energy gap is strongly anisotropic. have been introduced to describe materials in which the elec­ If this could be definitely established, it would be the culmina­ tronic states have a characteristic energy orders of magnitude tion of a long search for this phenomenon. smaller than in ordinary metals. If we write the energy t:( k) in There is no doubt that a proper understanding of this super­ a free-electron form (t:(k) = h 2 k2/ 2m*), then since the wave­ conducting state requires a clear understanding of the preceding vectors k of the electron determined by the interatomic spacing normal state with its remarkable properties. Early theories of are not much different, the effective mass m * must be orders of the properties of metals always assumed that, whereas the con­ magnitude larger than the free-electron value and in some cases duction electrons interact with the ionic lattice forming the solid, m* is a fair fraction of the proton mass. These materials are they do not interact at all among themselves. Quantum statistics intermetallic compounds in which one of the constituents is a then determine the low-temperature properties of this electron rare-earth or actinide atom, with partially filled 4f- or Sf-electron gas, two of which are of particular importance in ihe context shells. At high temperatures these materials behave as if the of our discussion. The specific heat of this electron gas cP varies f-electrons were localized on their atomic sites, as in conven­ linearly with temperature as T ~ 0 K (that is, cP = yT). The tional rare-earth and actinide compounds, where any itinerant low-temperature magnetic susceptibility, x, is independent of electrons are in states derived from loosely bound atomics-, p­ temperature. In this simple theory the ratio and d-orbitals. As the conventional materials are cooled, the x h =3µ.~ / w 2 k~ (1) atomic moments due to the f-electrons order spontaneously, mostly antiferromagnetically, less often ferromagnetically. By is obviously a universal number. µ.s = 9.27 x 10- 21 erg G- 1 is the contrast, in the heavy-electron systems some of the f-electrons Bohr magneton and ks = 1.38 x 10- 16 erg K- 1 is the Boltzmann become itinerant at low temperatures and form a metallic state constant. The factor which determines the magnitude of both x with the characteristics described above. and y is the density of electronic states per unit energy, N( EF). Recently the exciting discovery was made that in some of at the Fermi energy E F.( E F is the energy up to which all possible these new materials the heavy electrons form a superconducting states of the electron gas are occupied at T = 0 K.) Hence N( E F) 3 4 6 state at very low temperatures ' ' . Superconductivity in ordinary varies inversely with the characteristic energy of the electrons, metals is associated with an instability of itinerant electrons; leading N(EF) to be proportional to m *. This simple concept thus it was surprising that it should also occur where normal- is quite adequate to describe the qualitative low-temperature features of simple metals, for which y is of the order of 1 mJ 1 1 *Permanent address: Eidg. Tech. Hochschule, ETH-Hiinggerberg, mol- K- , x = 10-s e.m.u. mol- t and TF = EF/ ks, the Fermi CH-8093 Zurich, Switzerland. temperature, is -104-105 K. The experimental facts quoted in