This article originally appeared in The Skeptic, Volume 15, Issue 3, from 2002.
Many people worldwide report inexplicable experiences of apparitions, sounds, smells, sensed presences, bodily sensations, and physical manifestations that suggest paranormal origins (Haraldsson, 1985; Ross & Joshi, 1992). These phenomena are usually referred to as “haunts” when they seem tied to a particular location and as “poltergeists” when there is an outbreak of such phenomena in the presence of a specific person or persons (called a focal person or agent). All too often, the public is presented information on ghosts and poltergeists by self-styled “ghost hunters” or spiritualist teachers. In short, hauntings and poltergeists are a fertile breeding ground for quackery, deception, self-deception, frauds, and hoaxes.
We produced our edited book, Hauntings and Poltergeists: Multidisciplinary Perspectives (Houran & Lange, 2001) to provide a reliable anthology that discusses what is really known about these phenomena. The book has contributions by a variety of authors, each of whom has their own version of what constitutes hauntings and poltergeists — some take a believer’s perspective, some are sceptics. Regardless of one’s persuasion, we believe that it is instructive to see the case that can be made for each perspective.
In our own chapter we conclude that ghostly outbreaks tell us more about the living than the dead. That is, these experiences offer a unique opportunity to study how magical thinking, hallucination, suggestion, and responses to subtle environmental stimuli operate in unusual, but naturalistic, settings. As we have repeatedly argued (Lange & Houran, 1998, 1999, 2001a; Houran, Kumar, Thalbourne, & Lavertue, submitted), we believe that the study of haunts and poltergeists shares many features with other controversial phenomena such as contagious (mass) psychogenic illness, sick building syndrome, and other forms of alleged “environmental illness” (Lundberg, 1998).
The preceding does not make haunts and poltergeists any less mysterious or intriguing. Instead, it signifies hope that we can eventually understand and control these experiences, which often elicit deep, negative, and long-lasting effects on people’s lives. In short, we ignore all “for” and “against” debates, and rather try to determine what haunts and poltergeists may teach us about people. The following summarizes our work thus far.
Haunts and Poltergeists as Social Facts
Poltergeist-like episodes may be characterized as clustered perceptions of ambiguous psychological experiences and physical manifestations that often focus around adolescents, and particularly females, during periods of psychophysical stress. Similar sensory experiences and physical phenomena that persist over long periods of time at a particular location are called hauntings (Roll, 1977; Gauld & Cornell, 1979).
Hauntings and poltergeists have been reported in nearly all cultures throughout history, and their interpretation and the way people cope with them seem to be related to psychological, social, and cultural conditions (Roll, 1977; Hess, 1988). For instance, Carrington and Fodor (1951) collected incidents of “stone-throwing” poltergeists from as early as A.D. 530 up through the medieval period. Pliny the Younger (1751) wrote a report of a haunted house that came to his attention, and A. R. G. Owen (1964) published summaries of poltergeist-like episodes from the 19th and early 20th centuries.
Perhaps the first scientist to take these experiences seriously was the British physicist and chemist Robert Boyle. He encouraged a Protestant minister named Francis Perrault to publish a treatise on “the devil in Mascon,” which chronicled inexplicable noises and movements of objects that suddenly began in Perrault’s home in 1612 (Thurston, 1954). Modern cases have appeared in mainstream journals (e.g., Leon, 1975; Persinger, Koren, & O’Connor, 2001), but most are published in the parapsychological literature. For example, McHarg (1973, pp. 17-18) reported a relatively recent case in which a 13-year-old girl in the Midlands, England, began having hallucinations of people:
At first she saw an old man, who was taken to be her long-deceased grandfather. Then, in 1971, she repeatedly saw a young girl who claimed to have been strangled in 1808 and who wished to be buried in consecrated ground…Involvement of the rest of the family and of friends began when they witnessed ostensible poltergeist phenomena such as doors and curtains opening and shutting, objects moving… Apparitions… then began to be seen by others, both singly and collectively. These were not only of dead persons, but also of persons known to be alive. Dogs, bears, birds and devilish “horny things” were also seen – a coldness was usually experienced in the part of the body nearest to the apparition. Shared apparitions sometimes appeared, to different observers, to be differently dressed.
Although these types of experiences are depicted comically in movies like Ghostbusters and Beetlejuice, they often invoke fear and anxiety in real life (Rogo, 1974; Hufford, 1982). In fact, the development of the “Religious or Spiritual Problem” diagnostic code in the DSM-IV (see Turner, Lukoff, Barnhouse, & Lu, 1995), coupled with publications discussing how to address anomalous experiences in psychotherapy (e.g., Hoyt, 1980; Hastings, 1983; Kauffman, 1993; Peteet, 1994), reflects that “paranormal” experiences are increasingly recognized as legitimate issues in clinical psychology and psychiatry.
Why Do People Experience Haunting and Poltergeist Outbreaks?
Some argue that this question is not a topic for serious inquiry and readily dismiss accounts as fraud or symptoms of mental illness. Indeed, deliberate deceit (see e.g., Owen, 1964) and self-deception (see e.g., Eastham, 1988) can explain some cases, and experiences typical of haunting and poltergeist outbreaks are also commonly encountered in schizophrenia and affective disorder. However, Cox (1961) noted that some fraudulent cases also contain manifestations that do not seem manufactured. Moreover, psychopathology is implausible as a general explanation since up to 10% of the people in the general population worldwide have experienced a ghost or poltergeist (Haraldsson, 1985; Ross & Joshi, 1992).
It should not be surprising that no general consensus exists concerning the aetiology of hauntings and poltergeists. The most popular explanation for hauntings and poltergeists is that of discarnate agency (Stevenson, 1972), i.e., the notion that an organism can survive bodily death. Since the evidence offered in support of these survival points of view is generally ambiguous and subjective (Gauld, 1977), Stevenson (1972) argued that discarnate agency is evidenced only by phenomena reflecting intelligence and purpose – i.e., as opposed to manifestations such as explosive sounds or random movements of objects. Even if this reasoning is accepted, however, it appears that cases involving apparitions (and therefore presumably discarnate agents) were not significantly more likely to involve “intelligent or purposeful” manifestations (Alvarado & Zingrone, 1995).
Psychokinesis (PK), i.e., the idea that anomalies result from interactions between shifts in the consciousness of living human beings and the physical environment (Roll, 1977; Radin & Rebman, 1996), can be seen as a contemporary version of the above. Although some experimental evidence for PK is suggestive (e.g., Radin, 1997; Jahn, Dunne, Nelson, Dobyns, & Bradish, 1997), the putative effect sizes are far too small to account for the large-scale phenomena often reported during haunts and poltergeist-like episodes (flying objects, moving furniture, etc.).
It has been argued therefore that PK effects can be magnified under certain conditions (yet to be described), thereby producing dramatic physical phenomena. However, explanations invoking large-scale PK (also called macro-PK or recurrent spontaneous psychokinesis, RSPK) are moot until definitive evidence is presented.
Other researchers adhere to more conventional explanations. A nice summary is provided by Tandy and Lawrence (1998, p. 360) who listed
… water hammer in pipes and radiators (noises), electrical faults (fires, phone calls, video problems), structural faults (draughts, cold spots, damp spots, noises), seismic activity (object movement/destruction, noises), electromagnetic anomalies (hallucinations), and exotic organic phenomena (rats scratching, beetles ticking).
Interestingly, they also reported that standing air waves can elicit sensory experiences suggestive of ghosts, a natural cause which had not been documented previously.
As we have argued in several places (Lange & Houran, 1998, 1999, 2001a), there is a multitude of ambiguous psychological and physical phenomena which, given the proper context, can be interpreted as paranormal (for extensive discussions of ambiguous stimuli, see e.g., Houran & Lange, 1996c; Houran, 1997a; Houran & Lange, 1998). Like other authors (e.g., Cone, 1995; Maher, 1988), we argue that a more fruitful approach to establishing a general theory of haunts and poltergeists is to study the affective and cognitive dynamics of people’s interpretation of ambiguous stimuli. According to this view, ambiguous phenomena such as listed by Tandy and Lawrence become paranormal only after being interpreted as such against the backdrop of a shared human reality. In this sense, hauntings and poltergeist-like episodes are delusions. The remainder of this article outlines the body of empirical research that has led us to this conclusion. For additional information and more detailed discussions the reader is referred to Lange and Houran (2001a).
Modelling Poltergeist-like Episodes as Delusions
While performing content analyses of accounts of haunts and poltergeist-like experiences (Lange, Houran, Harte & Havens, 1996), we were struck by the fact that many percipients seemed eager to draw very definite and far-reaching conclusions from very limited and ambiguous information. Although percipients may insist that their experiences reflect consensual reality, it became increasingly clear that the contents and modalities of a wide variety of paranormal experiences are consistently and predictably shaped by “contextual variables” in the environment. This was true not only for haunts and poltergeists; similar conclusions were reached in analyses of angelic encounters (Lange & Houran, 1996), deathbed visions (Houran & Lange, 1997a), and anomalous photographic effects (Lange & Houran, 1997a). For a meta-analysis of these studies, see Houran (2000).
Subsequent research (Houran & Lange, 1996a, 1997b; Houran, 1997c) indicated that percipients’ experiences and beliefs concerning poltergeist-like phenomena also vary with their tolerance of ambiguity. Tolerance of ambiguity is an emotional and perceptual personality variable first described by Frenkel-Brunswick (1949). She described intolerance of ambiguity as the tendency to resort to black and white solutions characterized by premature closure, often at the neglect of reality. In essence, intolerance of ambiguity results in rapid and overconfident judgment of ambiguous stimuli or events. Budner (1962) considered tolerance of ambiguity as a motivating factor for individuals in social situations as well. In particular, he defined intolerance of ambiguity as the tendency to perceive ambiguous situations as threatening, whereas tolerance of ambiguity denoted the tendency to perceive ambiguous situations as desirable.
The question arose therefore how contextual variables and tolerance of ambiguity interact to guide percipients’ perceptions of the paranormal. Integration is obtained by approaching the perception of poltergeists and kindred phenomena within an attribution theory framework – i.e., such phenomena are seen as delusions involving an interaction among paranormal beliefs, paranormal experiences, and fear of the paranormal. We extended earlier views on this topic (Kihlstrom & Hoyt, 1988; Maher, 1988; Reed, 1988) by incorporating contextual variables, and factors like percipients’ age, gender, and tolerance of ambiguity into a single process model.
This model was first proposed in Lange and Houran (1998; cf. Lange & Houran, 1999. This diagram shows the nature and direction of the relations among these variables, as indicated by the directional arrows. For simplicity, only the signs (as indicated by the + and – labels), but not the magnitudes, of the various regression weights are shown. For instance, links of the form A + B indicate that variable A produces an increase in variable B, whereas links like A – B indicate that variable A inhibits variable B.
This model has been replicated in two separate studies reported in Lange and Houran (1998) and it is based on the responses of people who claimed to have had at least one haunting or poltergeist experience as originally defined in Lange et al. (1996, p. 757). Paranormal belief, experience, and fear were measured by the Anomalous Experiences Inventory (Kumar, Pekala & Gallagher, 1994), whereas tolerance of ambiguity was assessed by MacDonald’s (1970, p. 793) AT-20 scale. Although contextual variables are shown in Figure 1, they were not directly measured in the path analysis studies. Instead, their effects are inferred from the results of other studies (Lange et al., 1996; Lange & Houran, 1996; Houran & Lange, 1997a). To facilitate the presentation of the model, three types of variables are distinguished:
1. Contextual Variables
Just as the same ambiguous round shape can be perceived as an orange or a cup of water depending upon whether the person is hungry or thirsty (Horowitz, 1975), so can an apparition be perceived as a demon, an angel, or a deceased loved one, based on environmental influences such as embedded cues, demand characteristics, symbolic-metaphorical references, as well as an individual’s psychophysical state or prior beliefs (for a review, see Lange et al., 1996, pp. 755-758). We found that the more specific aspects of paranormal experiences are congruent with contextual variables as well. For instance, a person in an empty ballroom might hear the sound of waltz music and see people dancing, whereas a person in a room with a prominent lavender hue might report the smell of lilacs. Moreover, the number of experiential modalities increased with the number of contextual variables. That is, when one contextual variable is present percipients might only see an apparition, whereas the introduction of additional contextual variables might induce them to hear or smell something as well.
Most of our work on contextual variables involved content analyses of percipients’ personal accounts. However, it is not difficult to demonstrate context effects experimentally. For instance, in one study (Lange & Houran, 1997b) individual participants were taken on a tour of a performance theatre under renovation, and afterwards they completed an experiential questionnaire by Green, Parks, Green, Guyer, Fahrion and Coyne (1992, pp. 74-78). In the control condition the theatre was said to be undergoing remodelling, whereas in the “informed” condition the participant had been told that the theatre was haunted and to be aware of any unusual experiences. As expected, this manipulation elicited a considerably greater number of physical, extrapersonal, and transpersonal experiences relative to the control group. Since contextual variables guide the interpretation of ambiguous stimuli, Figure 1 shows a link from contextual variables to paranormal experiences. The link to paranormal beliefs was included because such beliefs are an integral part of the context. Our theatre experiment further showed that those in the “informed” condition experienced significantly more intense emotional responses, a finding that is consistent with the process described next.
2. Process Variables
Attribution theory explains delusions as a byproduct of a percipient’s failure to find a standard explanation for ambiguous stimuli. In particular, such theories assume that certain intense ambiguous experiences may be interpreted as personally significant to the perceiver, and that a lack of a consensual explanation for these experiences leads to the arousal of fear. Often, such fears can be reduced through essentially “normal” reasoning. However, if no standard explanation can be found then contextual variables may suggest a paranormal explanation. Once fear of the anomalous is reduced in this way, the resulting explanation persists as a defence mechanism, thereby becoming virtually immune to persuasion and counter-argument.
The above account is consistent with the finding of the negative-feedback loop shown in the top part of Figure 1. According to this figure, any increases (decreases) in belief and experience are accompanied by a corresponding decrease (increase) in fear, thereby neutralizing the initial changes. However, variations in the environment provide a continuous source of ambiguous stimuli in need of an interpretation, and thus a person’s beliefs and fears are unlikely to remain stable. Also, there is evidence that experiences, beliefs, and fears may be affected by self-reinforcing attentional biases at the social (Colligan, Pennebaker & Murphy, 1982; Wessely, 1987) as well as at the individual level.
For instance, in one study (Houran & Lange, 1996b) we asked a married couple to chronicle the frequency of strange or unusual events in their unhaunted residence for a period of thirty days. The observed events included complex and repeated movements of a particular object (a souvenir voodoo mask), erratic functioning of a telephone in the same general area of the house, and a mysterious voice. Although few anomalous events were observed over days 1 through 5, the frequency increased dramatically over days 6 through 15, only to die out thereafter. Since the distribution of events conformed to the logistic growth curve typical of infectious processes, we dubbed the effect “perceptual contagion”. Interestingly, research on real poltergeists also showed such clustering effects (Roll, 1968, 1969, 1970), thus suggesting that perceptual contagion may play a role in these cases as well (see Jones & Jones, 1994, p. 38).
Note that the negative-feedback loop in Figure 1 implies that beliefs directly induce experiences (and not vice-versa), and similar conclusions were reached in our research on death anxiety (Lange & Houran, 1997c). At first sight, these findings contradict Irwin’s (1993) review of the literature indicating that a direct link between experience and belief cannot be excluded. However, this contradiction may be more apparent than real. First, although experiences lead to a decrease in fear through belief, beliefs are also fuelled by ambiguity intolerance. Thus, the net effect of an increase in paranormal experiences on paranormal beliefs may be negligible. Second, the presence of contextual variables affects paranormal beliefs and experiences simultaneously and in a
similar fashion. It is not surprising, therefore, that belief and experience scales should show a positive correlation. Third, we believe that it is difficult, if not impossible, to induce paranormal beliefs without simultaneously introducing paranormal experiences, and vice-versa. Hence, the question whether beliefs or experiences are primary cannot really be answered. Instead, we prefer to interpret the circular relation between paranormal beliefs and experiences as characteristic of the delusional process.
3. Exogenous Variables
The bottom part of Figure 1 shows that paranormal experiences and beliefs are related to more stable characteristics of the percipients such as age, gender, and tolerance of ambiguity. That is, fear of the paranormal is greater for women, older people, and those with a low tolerance of ambiguity. Remember that anecdotal reports suggest that poltergeist-like experiences focus around young women (Owen, 1964; Roll, 1977; Gauld & Cornell, 1979). This notion is consistent with our finding that being female is associated with a greater fear of the paranormal as well as a lower tolerance of ambiguity. Unfortunately, our samples contained very few adolescents and no measures of psychological stress were obtained, and it was not possible to verify the focusing effect in its entirety. However, Keinan (1994) has already shown that intolerance of ambiguity is associated with enhanced magical thinking, especially during times of stress. By contrast, those with a greater tolerance of ambiguity solve problems through logical analysis and are more successful in coping with stress (Kuypers, 1972; Parkes, 1984).
Figure 1 further shows that belief in the paranormal is both positively and negatively affected by tolerance of ambiguity, suggesting the existence of two different underlying processes. For instance, those with a high tolerance of ambiguity might consider poltergeists (and other paranormal phenomena) to be no more than an interesting or amusing curiosity. Thus, rather than inducing fear, paranormal experiences might inspire feelings of admiration, awe, or wonder in those with a high tolerance of ambiguity, perhaps not unlike the experiences of some artists (Zausner, 1998).
By contrast, it would appear that a fear of the paranormal presupposes the perception of a clear boundary between what is “normal” and what is not. Hence, the attribution explanation summarized earlier should apply mainly to individuals with a low tolerance of ambiguity.
Such individuals face the choice between two conflicting alternatives: they can maintain their preferred “normal” view of the world while incurring an increase in fear, or they can avoid this fear at the cost of having to accept an undesirable paranormal explanation. Paranormal explanations might be undesirable for a variety of reasons. Not only could such explanations threaten one’s peace of mind, they might also lead to negative social consequences when expressed publicly.
