Why Do Ghosts Speak English? Key Points
- Across thousands of investigations, ghosts of every era and origin — Neolithic hunters, Roman legionaries, plague victims — answer in short, idiomatic, present-day English, and the pattern is so routine that investigators no longer notice it.
- Spiritualism’s rapping tables and planchettes gave way to EVP recorders, spirit boxes, and the Estes Method, but the practice remains the same: a human turning ambiguity into speech.
- A genuine medieval, early modern, Latin-speaking, or creole-speaking voice would be largely unintelligible to modern ears, yet what appears on playback is contemporary vocabulary and grammar with no period accent or syntax.
- Japanese ghosts answer in Japanese, Mexican ghosts in Spanish, Chernobyl’s in Russian or Ukrainian — a far more revealing finding, since it points to the listener rather than the dead.
- Auditory pareidolia and top-down speech restoration, amplified by expectation and priming, supply the words from the listener’s own monolingual lexicon, while radio bleed, the ideomotor effect, leading questions, group consensus, and selective archiving reinforce the result.
- Telepathic-translation defenses are unfalsifiable and xenoglossy claims collapse under linguistic scrutiny, so a controlled, independently transcribed response in a language no investigator knew would be decisive — and its continued absence, despite decades of cheap recorders, is itself the most informative finding.
Introduction
Somewhere in an old institution in Ohio, a man with a digital recorder asks a question into the dark, waits, and later plays back a burst of static that his team agrees says, “Get out,” or “I am cold.” A similar squad in a Welsh castle gets much the same. When he answers at all, the legionary said to be walking the walls of a Roman fort along Hadrian’s Wall speaks serviceable modern English, sometimes with an American accent. It’s so constant throughout thousands of investigations that it has become invisible to the individuals who are making it. Ghosts of all eras and locations, from Neolithic hunters to Confederate soldiers to victims of the plague, typically speak in the modern language of whoever controls the microphone, and that language is commonly English. This fact deserves grave consideration, because its constancy teaches us a tremendous lot about how spirit contact is actually formed, even if it tells us very little about the dead.
Defining the Ghost and Its Instruments
First, it helps to know what a ghost is supposed to be. The prevailing folk concept in Anglophone paranormal culture is that a ghost is the surviving consciousness or personality of a dead human being, kept in a place by trauma, attachment, perplexity, or unfinished business. Ghosts in this paradigm are generally assumed to be non-physical yet able to produce local physical effects. Cold spots, footfalls, moved objects, electromagnetic fluctuations, and, above all, sounds are all phenomena associated with ghosts. In their research of American ghost belief and ghost hunting, Baker and Bader (2014) claimed that the persistence of ghosts as cultural characters arises precisely from their liminality, their position between categories of living and dead, tangible and immaterial, and science and religion. This in-betweenness renders ghosts infinitely interpretable. They are very difficult to define, which is why the history of spirit connection largely revolves around various devices.
The instruments have changed, but the fundamental practice has not. Nineteenth-century spiritualism produced table rapping, the alphabet board, trance mediumship, and the planchette, all of which needed a human mediator to be able to convert ambiguity into speech (Faraday, 1853). Researchers in the twentieth century replaced the medium with the tape recorder and created electronic voice phenomena, or EVP, which are faint sayings believed to have been recorded in noise (Banks, 2001). Today’s technology is more complicated. There are full-spectrum cameras, structured light sensors, temperature loggers, and the spirit box, a receiver that rapidly sweeps the AM or FM band so that bits of broadcast audio and static blur together. Some groups now use the Estes Method, in which an investigator with a blindfold and headphones repeats aloud what the spirit box seems to say, while teammates (not audible to them) ask the questions (Stevens, 2024). The technology appears to be scientific. As Eaton (2020), who has spent years in the field with investigators, puts it, this technology is less a matter of measurement than of providing a shared framework for interpretation, a means by which small groups can negotiate their way toward an agreed-upon experience of a presence.
The Linguistic Problem: Ghosts Speak the Investigator’s Language
In that framework, the subject of language is rarely discussed. It ought to be. If the voice is that of a person who died in 1349, 1653, or 1780, the linguistic expectations are huge. Middle English is not comprehensible to the untrained modern ear. Early modern English features different vowels, different syntax, and a significantly different lexicon. A Roman auxiliary might speak Latin or the local vernacular, and an enslaved individual in the colonial era might speak a Creole or a West African language. But what you hear when it’s played back is nearly always brief, colloquial, contemporary speech: “help me,” “leave,” “yes,” “hello,” names and obscenities. The traditional accent is absent. And so is the historical syntax. The same pattern is true in non-Anglophone countries where the language is exchanged, which is the key point. Ghosts in Japan speak Japanese, ghosts in Mexico speak Spanish, and Chernobyl’s spirits speak Russian or Ukrainian. The dead do not seem to speak English but rather the investigator’s own language, which is a much more intriguing and illuminating claim.
Pareidolia, Priming, and the Manufacture of Voices
There are several convergent reasons for this phenomenon, and most of them are from standard cognitive research, not parapsychology. The most powerful is auditory pareidolia, the perception of meaningful speech in ambiguous sound (Nees & Phillips, 2015; Williams & Blagrove, 2022). Human speech perception is aggressively top-down. We repair missing phonemes, fill gaps, and impose lexical order on degraded information. That is precisely what we need to do in regular life. In Deutsch’s (2003) phantom word demonstrations, listeners hear some personally meaningful words in looping nonsense quickly. Most EVP can be explained by psychoacoustics alone, stated Banks (2001) in his “Rorschach audio” paper. Anticipation critically influences what listeners perceive. Nees and Phillips (2015) played clips of EVPs, degraded speech, real speech, and noise to participants. Participants who were informed that the study focused on paranormal voices heard many different voices, and when they listened to one, they had a very low level of agreement on what it said. Williams and Blagrove (2022) partially replicated and obtained a comparable priming effect. The listeners are English speakers; hence, the lexicon accessible for restoration is in English. The perceptual machinery that produces the speech is monolingual. So the ghost speaks English.
The equipment emphasizes the phenomenon. Kansas radio is in English. Therefore, any identifiable word produced by a spirit box sweeping the FM spectrum in Kansas was almost certainly in English before any spirit interacted with it. Even friendly ghost hunters accept that stray transmission is the primary problem. The ad hoc criteria for ruling it out, such as interpreting obscenity as proof of genuineness, are not remotely dependable (Stevens, 2024). Add to the mix the ideomotor effect, described in the nineteenth century by Carpenter (1852) and by Faraday (1853), which interprets planchette and dowsing answers as unconscious muscular movement, and you have a second channel by which the investigator’s own linguistic knowledge enters the record. You hear that? That’s that thing, right there. This includes the social dynamics documented by Eaton (2020), where the suggestions made by a senior member regarding what a clip says become the consensus of the group. Then there is selection: incomprehensible clips are eliminated, while intelligible ones are archived, captioned, and broadcast, priming television audiences with subtitles before they even hear the audio. Finally, remember that the whole practice is an Anglophone cultural export, promoted throughout the world by American and British programs, which takes with it its ideas about how spirits behave (Baker & Bader, 2014).
Counterarguments and the Test of Xenoglossy
Believers have presented counterarguments, and these are fairly stated. Some have suggested that spirits communicate telepathically, with the message translated by the percipient’s own mind into their native tongue. This notion has a Swedenborgian heritage and was explored at length by nineteenth-century Spiritualists (Edmonds & Dexter, 1855). The others claim the dead learn, observe, and adapt. Both ideas are unfalsifiable as they stand, and both concede, quietly, the fact that there is no independent linguistic substance. The best we can do is something tested—xenoglossy, the claim that someone is actually speaking an unlearned language. Stevenson (1974, 1984) has claimed that there are a few such occurrences, but they have been reanalyzed by linguist Sarah Thomason, who has found that the evidence is too weak to back up the assertion (Thomason, 1984, 1987). Historical mediumistic xenoglossy does even worse. Analysis shows that the angelic Enochian of Edward Kelley is based upon English (Laycock, 2001).
Conclusion
Ghosts speak English to English-speaking ghost investigators because clouds look like whatever the viewer was contemplating recently. The voices do not come from the past but are constructed at the point of reception from the investigator’s own vocabulary, expectations, questions, radio spectrum, and social setting, and then stabilized through group agreement and edited for broadcast (Banks, 2001; Eaton, 2020; Nees & Phillips, 2015). This phenomenon is not a charge of fraud, nor does it need anyone engaged to be silly; the perceptual processes involved are perfectly normal, useful, and generally automatic (Carpenter, 1852; Williams & Blagrove, 2022). It does, however, provide a defined standard to the field. Extraordinary evidence would be an authentic reply utterance in a language unknown to any of the investigators, captured under controlled conditions and transcribed independently by a linguist with no knowledge of the case, and it would immediately attract scientific attention (Thomason, 1987). The most revealing conclusion the ghost hunting community has produced is that such recordings are notably lacking, even after decades of effort and cheap recording technology in millions of hands.
References
Baker, J. O., & Bader, C. D. (2014). A social anthropology of ghosts in twenty-first-century America. Social Compass, 61(4), 569–593. https://doi.org/10.1177/0037768614547337
Banks, J. (2001). Rorschach audio: Ghost voices and perceptual creativity. Leonardo Music Journal, 11, 77–83. https://doi.org/10.1162/096112101750343178
Carpenter, W. B. (1852). On the influence of suggestion in modifying and directing muscular movement, independently of volition. Proceedings of the Royal Institution of Great Britain, 1, 147–153.
Deutsch, D. (2003). Phantom words and other curiosities [Audio CD]. Philomel Records.
Eaton, M. A. (2020). Sensing spirits: Paranormal investigation and the social construction of ghosts. Routledge.
Edmonds, J. W., & Dexter, G. T. (1855). Spiritualism (Vol. 2). Partridge & Brittan.
Faraday, M. (1853, July 2). Professor Faraday on table-moving. The Athenaeum, (1340), 801–803.
Laycock, D. C. (2001). The complete Enochian dictionary: A dictionary of the angelic language as revealed to Dr. John Dee and Edward Kelley. Weiser Books.
Nees, M. A., & Phillips, C. (2015). Auditory pareidolia: Effects of contextual priming on perceptions of purportedly paranormal and ambiguous auditory stimuli. Applied Cognitive Psychology, 29(1), 129–134. https://doi.org/10.1002/acp.3068
Stevens, H. (2024, August 30). The Estes method is an update to the ghost hunter’s spirit box, with all the same flaws. The Skeptic. https://www.skeptic.org.uk/2024/08/the-estes-method-is-an-update-to-the-ghost-hunters-spirit-box-with-all-the-same-flaws/
Stevenson, I. (1974). Xenoglossy: A review and report of a case. University Press of Virginia.
Stevenson, I. (1984). Unlearned language: New studies in xenoglossy. University Press of Virginia.
Thomason, S. G. (1984). Do you remember your previous life’s language in your present incarnation? American Speech, 59(4), 340–350. https://doi.org/10.2307/454785
Thomason, S. G. (1987). Past tongues remembered? Skeptical Inquirer, 11(4), 367–375.
Williams, J. M., & Blagrove, M. (2022). Paranormal experiences, sensory-processing sensitivity, and the priming of pareidolia. PLOS ONE, 17(9), e0274595. https://doi.org/10.1371/journal.pone.0274595
Leave a Reply