Telepathy

Abstract

Telepathy is a form of parapsychology: the claimed direct transmission of thoughts, images, or feelings from one mind to another without any known sensory channel. It is one of the purported extrasensory abilities, distinguished from clairvoyance, the claimed perception of distant objects, and precognition, the claimed perception of future events. For cognitive psychology the interest is not the phenomenon, which has no accepted evidence or mechanism, but the century-long research program built to test it: the forced-choice card guessing of the 1930s, the ganzfeld sensory-reduction procedure, and the meta-analytic debate over whether small above-chance hit rates reflect a real signal or the artifacts of flawed method. That program produced some of psychology's most instructive methodological controversies, and one experiment on a related psi claim helped trigger the replication crisis.

Keywords: telepathy, extrasensory perception, ganzfeld, psi, parapsychology

Telepathy is the alleged ability of one person to become aware of another's mental contents, their thoughts, images, or emotions, without communication through sight, sound, or any recognized sense. The word, coined in 1882 from the Greek roots for distant feeling, names a claim that has never been established under controlled conditions, yet the effort to test it rigorously has occupied experimental psychologists for the better part of a century. That effort is the reason a construct with no accepted empirical support earns a place in a cognitive-science reference: telepathy is the paradigm case for asking how a small, statistically significant departure from chance should be interpreted, and the debate over it sharpened the tools, meta-analysis, preregistration, and the critique of flexible analysis, that the wider field now uses to police its own claims (Cardeña, 2018; Hyman, 1994).

What Telepathy Is

Telepathy is defined as the acquisition of information about another person's mental state by means other than the recognized senses or inference. It is one member of a family of claimed abilities collectively termed psi, the parapsychological label for anomalous information transfer, and it is separated from its siblings by the source of the information rather than by any difference in mechanism. In clairvoyance the information is held to come from a physical object or scene; in precognition it is held to come from a future event; in telepathy it is held to come from another mind. Because a well-designed experiment can rarely isolate a living sender's thoughts from the surrounding physical state of the world, telepathy and clairvoyance are in practice difficult to distinguish, and much of the literature treats them together under the broader heading of extrasensory perception, the term Rhine established for the whole class (Rhine, 1934).

The scientific study of telepathy began not with an attempt to prove it but with an attempt to measure it against chance. Rhine's card-guessing experiments at Duke University in the 1930s used a deck of five distinct symbols and asked a percipient to name each card in turn; with five equally likely alternatives, the expected hit rate under pure guessing is exactly one in five, and the whole evidential question reduces to whether an observed rate departs from 0.20 by more than sampling variation would allow. This forced-choice design is the paradigm's foundation: it converts a vague claim of mind-reading into a binomial hypothesis test with a known null, so that the burden falls entirely on demonstrating a reliable, replicable excess over the baseline. Figure 1 shows the claimed channel that such tests are built to detect, and the first demonstration lets a forced-choice run be simulated against its chance baseline.

Figure 1

The Claimed Telepathic Channel

Testing Telepathy: The Ganzfeld

Forced-choice card guessing dominated the first decades of research but proved vulnerable to a battery of sensory-leakage and recording problems, and by the 1970s the field had largely moved to the ganzfeld procedure, which remains the flagship telepathy paradigm. The word is German for whole field, and the method rests on the hypothesis that a weak psi signal is drowned out by ordinary sensory input and might surface if that input is reduced. A receiver reclines in a state of homogeneous, unpatterned stimulation, halved translucent spheres over the eyes bathed in red light, white noise in the ears, while a sender in a distant room concentrates on a randomly chosen target, usually a video clip or image. The receiver reports a running stream of impressions, and afterward judges which of four candidate clips, one target and three decoys, best matches those impressions.

This four-alternative judgment sets the chance baseline at exactly 25%, the number against which every ganzfeld result is read. The autoganzfeld systems developed by Honorton in the 1980s automated target selection, presentation, and randomization to close the sensory-leakage and experimenter-handling loopholes that had compromised earlier work, and it was the pooled autoganzfeld database that Bem and Honorton brought to a mainstream audience, reporting a hit rate near 32%, several points above the 25% baseline and, across the accumulated trials, conventionally significant (Bem & Honorton, 1994). Whether that excess is a signal or an artifact is the question the next section takes up; the second demonstration lets a ganzfeld series be run so the sampling behavior of the hit rate around its 25% baseline can be seen directly.

Evidence and Critique

The evidential history of telepathy is a sustained argument between meta-analysis and methodological critique, and it is the most disciplined instance of that argument in psychology. On the affirmative side, statistician Jessica Utts argued that the cumulative parapsychological database yields effect sizes that are small but consistent and non-zero, and that by the ordinary standards applied elsewhere in the behavioral sciences the anomaly would be considered replicated (Utts, 1991). A later meta-analysis of free-response studies from 1992 to 2008 reported a reliable above-chance effect and framed it as support for the noise-reduction model behind the ganzfeld (Storm et al., 2010), and a recent narrative review of the whole experimental literature concluded that the accumulated evidence warrants serious scientific attention rather than dismissal (Cardeña, 2018).

The critical side does not dispute that the pooled numbers exceed chance; it disputes what produces the excess. Hyman's analysis of the ganzfeld database identified a relationship between methodological quality and outcome, with the strongest effects concentrated in the studies with the weakest controls, and argued that flexible stopping rules, multiple analysis paths, and selective reporting could generate the observed excess with no psi at all (Hyman, 1994). The decisive empirical blow was a multi-laboratory attempt to reproduce the autoganzfeld result: pooling new, well-controlled studies, Milton and Wiseman found a hit rate indistinguishable from chance, breaking the replication the earlier meta-analysis had claimed (Milton & Wiseman, 1999). Modeling work later showed how a modest rate of questionable research practices alone could manufacture a significant meta-analytic effect in this exact literature (Bierman et al., 2016).

The controversy reached the wider field through a related psi claim. Bem reported nine experiments purporting to show precognition, that participants' responses were influenced by stimuli presented after the response, and published them in a flagship social-psychology journal (Bem, 2011). The paper used entirely standard methods and analyses, which was precisely its impact: if orthodox practice could produce significant support for an impossible effect, the practice, not the effect, was the problem. A Bayesian reanalysis showed the data did not support the psi hypothesis once appropriate priors and the flexibility of the analyses were accounted for (Wagenmakers et al., 2011), and independent laboratories failed to replicate the retroactive-recall effect (Ritchie et al., 2012). Table 1 sets out the principal positions in the debate, and the third demonstration shows how analytic flexibility can push a null result across the significance threshold.

Running beneath the evidential dispute is a psychological question the critical literature answers with more confidence than the affirmative one answers its own: why belief in telepathy is so widespread when the controlled evidence is so weak. The misattribution hypothesis holds that people who believe in psychic ability tend to misremember and reinterpret ordinary events, treating chance coincidences, such as thinking of a person just before they call, as evidence of a mental link while forgetting the far more numerous times no such coincidence occurred (Wiseman & Watt, 2006). Individual differences in reasoning style predict the belief: reliance on intuitive rather than analytic processing, and susceptibility to probabilistic reasoning biases, are associated with stronger endorsement of paranormal claims including telepathy (Williams et al., 2022).

Worked Example

Consider the ganzfeld baseline the second demonstration uses, and make the evidential logic precise. In a four-alternative ganzfeld trial the receiver judges one target against three decoys, so under pure guessing the probability of a hit is exactly 0.25. Suppose a series of 100 trials yields 32 hits, the kind of rate the autoganzfeld database reported. Is 32 a real excess or ordinary sampling noise?

Under the null of chance guessing the number of hits is binomial with n = 100 and p = 0.25, so the expected count is np = 25 and the standard deviation is the square root of np(1 - p), that is the square root of 100 times 0.25 times 0.75, which is the square root of 18.75, or about 4.33. The observed 32 hits sits 32 minus 25, or 7, hits above expectation, which is 7 divided by 4.33, roughly 1.6 standard deviations above the mean, a one-tailed p of about 0.05. A single such series is therefore only marginally suggestive, exactly at the threshold, and would not by itself persuade anyone.

The lesson the paradigm turns on is what happens when many such marginal series are pooled. Combining independent trials shrinks the standard error in proportion to the square root of the total count, so if the same 32% rate held across 1,000 trials the excess would be about 5 standard deviations above chance and highly significant. This is why the entire dispute is meta-analytic: no single ganzfeld study settles anything, and the whole question becomes whether the pooled excess reflects a genuine constant effect or the accumulation of small biases, selective inclusion, and flexible stopping across a heterogeneous body of studies (Hyman, 1994; Utts, 1991). The arithmetic that makes pooling powerful is the same arithmetic that makes it fragile to even a low rate of questionable practice.

Discussion

Telepathy occupies an unusual position in the science of mind: a construct with no accepted evidence and no candidate mechanism that has nonetheless been more rigorously tested than many accepted effects, and whose testing improved the discipline that rejects it. The affirmative and critical camps do not, in fact, disagree about the raw numbers; they disagree about the inference from a small pooled excess to a real cause. The proponents apply the meta-analytic conventions of the behavioral sciences and conclude that a consistent, non-zero effect size means a real phenomenon. The critics argue that those very conventions, significance testing on flexibly analyzed data with selective reporting, are capable of manufacturing a stable small effect out of nothing, and that the failure of controlled multi-laboratory replication is the decisive datum a meta-analysis of the older literature cannot override.

The lasting significance of the telepathy program is therefore methodological rather than substantive. The ganzfeld debate was an early, unusually explicit rehearsal of the argument that later engulfed the whole of psychology: that a literature of individually significant, underpowered studies with undisclosed analytic flexibility can sustain belief in an effect that better-controlled work cannot reproduce. When Bem's precognition paper applied impeccably standard methods to an impossible claim and got a publishable result, it functioned as a reductio of those methods, and the reforms that followed, preregistration, direct replication, Bayesian reanalysis, and attention to questionable research practices, are in part the wider field learning the lesson the parapsychology critics had been pressing for decades. Telepathy is best understood in cognitive psychology not as an open empirical question but as the sharpest available case study in how to tell a real effect from an artifact of method.

Current Directions

The most consequential recent development is the arrival of large, fully preregistered, adversarial replication into the field. The Transparent Psi Project mounted a high-powered, multi-laboratory, preregistered replication of Bem's precognition experiment with the analysis pipeline fixed in advance and the data collection monitored, and found no evidence of the anomalous effect, demonstrating that the reforms born from the replication crisis can be applied to the very claims that helped provoke it (Kekecs et al., 2023). Alongside it, the modeling of questionable research practices has become quantitative: rather than merely asserting that flexible analysis could explain the meta-analytic signal, recent work estimates how large a signal a plausible rate of such practices would produce, and shows it is comparable to the effect parapsychology reports (Bierman et al., 2016). A third current is the continued case, still contested, that when the whole experimental corpus is assessed by standard meta-analytic criteria it should be treated as an open scientific question rather than a closed one (Cardeña, 2018). Running through the cognitive literature is a sharpening account of paranormal belief itself, linking endorsement of telepathy to specific, measurable biases in probabilistic reasoning and intuitive thinking style rather than to a global deficit in intelligence (Williams et al., 2022).

Common Misconceptions

Statistically significant ganzfeld results prove telepathy exists.

A small pooled excess over chance is consistent with a real effect but also with study-quality artifacts, selective reporting, and flexible analysis, and the effect fails well-controlled multi-laboratory replication. Significance is necessary, not sufficient, and does not by itself identify a cause (Hyman, 1994; Milton & Wiseman, 1999).

Scientists reject telepathy out of prejudice rather than evidence.

The construct has been tested for nearly a century with progressively tighter controls, and the rejection rests on replication failure and the demonstrable sufficiency of ordinary artifacts to produce the reported effects, not on a refusal to look (Bierman et al., 2016; Wagenmakers et al., 2011).

Vividly accurate personal experiences are evidence of a telepathic link.

The misattribution hypothesis explains such experiences as selective memory for coincidences: the striking hit of thinking of someone before they call is remembered while the vastly more frequent misses are not, and reasoning biases convert chance into apparent connection (Wiseman & Watt, 2006).

Glossary

Autoganzfeld.

An automated ganzfeld system in which target selection, presentation, and randomization are computer-controlled to remove the sensory-leakage and experimenter-handling loopholes of earlier manual procedures.

Clairvoyance.

The claimed extrasensory perception of distant or hidden physical objects and scenes, distinguished from telepathy by having a physical rather than a mental source of information.

Decline effect.

The tendency of a claimed parapsychological effect to shrink as procedures are tightened and studies are repeated, a pattern critics read as the removal of artifacts rather than the fading of a real ability.

Effect size.

A standardized measure of the magnitude of a departure from chance, the quantity meta-analyses pool across studies and the focus of the dispute over whether the parapsychological effect is real.

Extrasensory perception (ESP).

The umbrella term Rhine established for the acquisition of information without the recognized senses, comprising telepathy, clairvoyance, and precognition.

File-drawer problem.

The bias introduced when statistically significant studies are published while null results remain unpublished, inflating any meta-analytic estimate built from the literature that survives.

Forced-choice test.

A psi task in which the target is one of a fixed, small set of alternatives, giving a known chance baseline against which an observed hit rate is tested, as in Rhine's five-symbol card guessing.

Free-response test.

A psi task in which the receiver reports open-ended impressions and afterward judges which of several candidate targets they best match, as in the ganzfeld's four-alternative decision.

Ganzfeld.

A sensory-reduction procedure, from the German for whole field, that exposes a receiver to homogeneous visual and auditory stimulation on the hypothesis that reducing ordinary input lets a weak psi signal surface.

Meta-analysis.

The statistical pooling of many studies to estimate a common effect; the central tool of the affirmative case for psi and the central target of the critique that flexible practice can bias the pool.

Misattribution hypothesis.

The account that belief in psychic ability arises from selectively remembering coincidences and reinterpreting ordinary events, rather than from any genuine anomalous experience.

Precognition.

The claimed extrasensory perception of future events; the psi ability tested in Bem's 2011 experiments whose standard-method significance helped trigger psychology's replication crisis.

Psi.

The parapsychological cover term for anomalous information transfer or influence, encompassing the extrasensory abilities of telepathy, clairvoyance, and precognition and the influence claim of psychokinesis.

Questionable research practices.

Flexible analytic choices such as optional stopping, selective reporting, and undisclosed multiple comparisons that can inflate the rate of significant findings without any real effect.

Receiver.

The participant in a telepathy test who attempts to apprehend the target, also called the percipient; the person whose hit rate against chance constitutes the datum.

Sender.

The participant, also called the agent, who views or concentrates on the target and whose mental state the receiver is claimed to apprehend.

Telepathy.

The claimed direct transmission of thoughts, images, or feelings from one mind to another without any recognized sensory channel; the extrasensory ability whose information source is another mind.

Key Researchers

Daryl J. Bem (b. 1938). Professor emeritus of psychology at Cornell University; his 1994 autoganzfeld meta-analysis with Honorton brought ganzfeld telepathy claims to a mainstream audience, and his 2011 precognition report became a catalyst for psychology's replication crisis. Wikipedia - Wikidata - Faculty Page

Etzel Cardeña (b. 1957). Thorsen Professor of Psychology at Lund University and director of its parapsychology research unit; he authored a 2018 review in a flagship journal arguing that the cumulative experimental evidence for psi merits serious scientific attention. Wikipedia - Wikidata

Charles Honorton (1946-1992). American parapsychologist who designed the ganzfeld and automated autoganzfeld ESP protocols at the Psychophysical Research Laboratories and co-authored the 1994 meta-analysis that defined the modern telepathy debate. Wikipedia - Wikidata

Ray Hyman (b. 1928). Professor emeritus of psychology at the University of Oregon and a founder of organized scientific skepticism; his methodological critique of the ganzfeld database is the canonical statement of the artifact case against telepathy. Wikipedia - Wikidata - Faculty Page

Joseph Banks Rhine (1895-1980). Botanist-turned-parapsychologist at Duke University who founded modern experimental parapsychology and coined the term extra-sensory perception in his 1934 monograph, establishing the forced-choice card-guessing paradigm. Wikipedia - Wikidata

Richard Wiseman (b. 1966). Professor of the Public Understanding of Psychology at the University of Hertfordshire; his multi-laboratory non-replication of the autoganzfeld and his work on the misattribution of psychic belief anchor the empirical case against telepathy. ORCID - Wikipedia - Wikidata - Faculty Page

Frequently Asked Questions

What is telepathy?
Telepathy is the claimed direct transmission of thoughts, images, or feelings from one mind to another without any recognized sensory channel. It is one of the extrasensory abilities studied under parapsychology, distinguished from clairvoyance and precognition by having another person's mind as its supposed source of information (Rhine, 1934).

Is telepathy scientifically proven?
No. There is no replicable controlled evidence for telepathy and no accepted mechanism. Pro-psi meta-analyses report small above-chance hit rates, but these fail well-controlled multi-laboratory replication and can be explained by methodological artifacts and flexible analysis (Milton & Wiseman, 1999; Hyman, 1994).

What is the ganzfeld experiment?
The ganzfeld is the flagship telepathy paradigm. A receiver rests under homogeneous red light and white noise while a sender concentrates on a randomly chosen target, and the receiver then judges which of four candidates matches their impressions. The four alternatives set a 25% chance baseline against which the hit rate is tested (Bem & Honorton, 1994).

Why do proponents and critics disagree if they use the same data?
They agree the pooled hit rate slightly exceeds chance but disagree about the cause. Proponents read the small consistent effect size as a real phenomenon; critics argue that study-quality artifacts, selective reporting, and flexible analysis can manufacture such an excess, and that controlled replications return chance-level results (Utts, 1991; Bierman et al., 2016).

How did telepathy research affect the replication crisis?
Bem's 2011 report used entirely standard methods to obtain significant support for precognition, an impossible effect. This functioned as a demonstration that orthodox practice could produce false positives, and the Bayesian reanalyses and failed replications that followed helped drive the reforms of the replication crisis (Bem, 2011; Wagenmakers et al., 2011).

Why do so many people believe in telepathy?
The misattribution hypothesis holds that belief arises from selectively remembering coincidences, such as thinking of someone just before they call, while forgetting the far more numerous misses. Reasoning biases and an intuitive rather than analytic thinking style predict stronger paranormal belief (Wiseman & Watt, 2006; Williams et al., 2022).

What is the difference between telepathy and clairvoyance?
Both are claimed forms of extrasensory perception, but they differ in the source of the information. Telepathy is held to draw on another person's mind, while clairvoyance is held to draw on a physical object or scene. Because experiments rarely isolate a sender's thoughts from the physical world, the two are often tested together (Rhine, 1934).

Have recent, rigorous studies found any evidence?
The most rigorous recent test, a large preregistered multi-laboratory replication of Bem's precognition experiment with its analysis fixed in advance, found no anomalous effect. It showed that the reforms produced by the replication crisis can be applied to psi claims, and it returned a null result (Kekecs et al., 2023).

References