The Word and the Tradition
The Tibetan word sprul-pa, transliterated in the anglophone literature since the 1930s as tulpa, denotes what the Tibetan Buddhist and Bön traditions understood as a consciously-created emanation — a being brought into existence through the sustained focused visualisation of an advanced practitioner and, under specific and demanding conditions, endowed with sufficient apparent autonomy that others could perceive it, that it could perform actions its creator had not directly willed, and that it could persist in operation after the creator’s attention had been withdrawn. The classical Tibetan treatment of the practice occurs in the context of deity yoga — the central meditational technique of Vajrayāna Buddhism in which the practitioner, over periods of months or years, holds the detailed visualisation of a specific enlightened being until that being becomes perceptually and emotionally present to a degree that the practitioner’s ordinary categorical distinction between imagination and perception begins to break down. The deity yoga practices are not ordinarily described by Tibetan teachers as tulpa creation; they are described as the cultivation of an already-existing awakened quality by means of its iconographic embodiment, and the ontological question of what the deity is at the end of the practice is answered within a sophisticated theoretical framework that does not map cleanly onto the Western distinction between subjective and objective.
What the word tulpa did not mean in its original usage was a benign imaginary companion reliably accessible to casual practice. It meant the product of an advanced operation, conducted within a framework of traditional authority and with the supervision of a qualified lineage teacher, whose outcomes included the possibility of the emanation taking on characteristics the practitioner had not intended and had, in some recorded cases, considerable difficulty removing from operation once the practice was discontinued. The most widely cited Western account of the practice is that of Alexandra David-Néel, the French Buddhist practitioner and explorer whose Magic and Mystery in Tibet (1929) described in plain prose her own attempt, undertaken during an extended stay in a Himalayan retreat, to create a tulpa in the form of a jovial, harmless monk. The emanation developed as David-Néel described the practice of creating it developing: through sustained daily visualisation over a period of months, the monk became increasingly perceptually present, was eventually seen by third parties who had not been informed of the experiment, acquired expressions and postures David-Néel had not directly visualised, and — when she attempted to withdraw the practice — proved reluctant to dissolve. David-Néel reported that the dissolution process took approximately six months of sustained effort, that during this period the monk’s character shifted from the jovial form she had created toward a more disquieting one, and that she came to understand in personal terms what her Tibetan teachers had meant by warning that tulpa creation was a practice the lineage treated with appropriate caution. The account has been cited with suspicion in some parts of the academic literature because David-Néel was writing for a popular audience and had incentives to dramatise her experiences. It has been cited with considerable interest in other parts of the literature because the details of her account converge with the details of contemporary reports produced by practitioners who could not have been reading David-Néel.
The Online Rediscovery
In 2009, on a now-archived thread in the bodymind board of the image-board community 4chan, a user describing himself as a practitioner of what he called a Tibetan meditation technique began posting descriptions of his experiences creating what he called a tulpa — an autonomous internal companion with its own thoughts, voice, emotional register, and preferences. The thread attracted a cluster of interested readers, several of whom began their own practice experiments, and over the following two years the community documenting its own attempts at the practice migrated to dedicated forums, eventually settling at tulpa.info and the tulpas subreddit, where it has remained continuously active ever since. What had begun as a single thread on a pseudonymous image board became, over the course of roughly a decade, a subculture of several thousand practitioners who describe themselves as engaged in a shared practice and a body of practice-experience reports that, taken as a whole, constitute the largest and most methodologically self-conscious corpus of data on deliberate autonomous-entity cultivation assembled in any Western context.
The tulpamancy subculture, as it came to be called, differed from the traditional Tibetan context in essentially every respect that would ordinarily be considered relevant to the success of such a practice. The participants were predominantly young, anglophone, male, and without formal training in any contemplative tradition. The instruction they followed was self-assembled from a mixture of David-Néel, online summaries of Tibetan deity yoga, and the practice guides generated by earlier members of the subculture itself. There was no lineage, no qualified supervisor, no traditional framework, no commitment to Buddhist ethics or cosmology, and in most cases no prior meditative experience of any kind. The practitioners were frequently first drawn to the practice by the simpler motivation of creating a reliable imaginary companion — often, in the early years of the subculture, modelled on a character from an animated television programme. What they found, if the self-reports of the subculture are taken at face value and if the convergence of those reports is treated as meaningful, was that the practice worked. Not in every case. Not for every practitioner. But often enough, and with sufficient structural consistency across practitioners who had no contact with one another, that the phenomenon demanded more careful consideration than the initial dismissal as wishful thinking or role-play could accommodate.
The convergent report the subculture produced was, in summary, the following. The practice begins with sustained visualisation of the intended form, typically for periods of thirty minutes to several hours per day over a period of weeks or months. The practitioner holds the visualisation with intention and speaks to the envisioned entity as though it were already present, attending to responses whether or not any are perceived. For most practitioners, the early weeks produced nothing beyond what the practitioner had imagined. At some point, generally between two weeks and three months of sustained practice, a shift occurred: the responses from the envisioned entity began to arrive without the practitioner’s conscious effort to generate them. The responses carried affective tone, verbal content, and eventually what practitioners described as personality features the practitioner had not chosen and did not control. In later stages the tulpa began to speak independently of the practitioner’s prompting, to express preferences the practitioner had not suggested, to offer observations about the practitioner’s behaviour and emotional life that the practitioner had not been thinking, and in some cases to manifest as a visual presence accessible to the practitioner’s open-eyed perception. The stabilised tulpa, for practitioners who had reached what the subculture called forcing maturity, functioned as what orthodox psychology would describe as an autonomous internal agent and what the tradition David-Néel was describing would call a successfully created emanation. The experience of being host to a stable tulpa was described by practitioners in terms that did not resemble the experience of having an imaginary friend. It resembled the experience of sharing cognitive space with another person.
The Methodological Status
The tulpamancy literature exists almost entirely in the form of self-reports by practitioners with no training in research methodology, no access to independent verification, and no protection from the ordinary biases that affect self-reporting in any domain. A responsible reader should take this seriously. The reports are not laboratory data. They are accounts produced by people who have invested weeks or months in a practice whose success would confirm something they had committed to, and whose failure would embarrass them. The ordinary explanation of the reports — that the practitioners are, consciously or unconsciously, role-playing the expected outcome and mistaking the role-play for genuine autonomous emergence — is available and cannot be decisively refuted on the evidence the subculture itself has produced. What can be said is that the reports are more consistent, more detailed, and more structurally self-critical than a naive role-play hypothesis would predict, and that the practitioners themselves have been remarkably willing to distinguish between what they describe as successful tulpa operation and what they describe as their own fantasy — a distinction that would be unstable in a pure role-play account but that practitioners maintain with surprising consistency.
The most serious empirical study of the tulpamancy community to date is that of Samuel Veissière, an anthropologist and cognitive scientist at McGill, whose 2016 paper in Transcultural Psychiatry examined the phenomenon through semi-structured interviews with over thirty active practitioners combined with ethnographic observation of the online community. Veissière’s methodological framework was drawn from the anthropology of altered mental states and from the cognitive science of what he called sociocognitive phenotypes — the hypothesis that culturally-specific practices can reliably produce particular mental configurations by engaging ordinary cognitive capacities in unusual ways. His conclusions were cautious. He found that the practitioners’ reports were internally consistent, that the phenomenology was reliably reproduced across practitioners, and that the experiences were not plausibly explained by any dominant psychopathological framework. The tulpamancers were not, as a group, suffering from dissociative identity disorder in the clinical sense; they did not meet the criteria, they did not report the distress characteristic of that condition, and the experiences they described were experienced as additions to rather than disruptions of their cognitive and emotional lives. Veissière proposed that the phenomenon was best understood as a form of culturally-constructed deliberate dissociation that used ordinary cognitive capacities — auditory imagination, sustained attention, imaginative rehearsal — to produce a stable autonomous internal agent that the practitioner’s conscious self interacted with as though with another person.
This reading is compatible with the materialist production model, and it has the virtue of remaining within the framework the mainstream cognitive science literature is prepared to operate in. It treats the tulpa as real in the sense that the practitioner’s experience of it is real — as real as any thought or feeling — without committing to any stronger ontological claim about the tulpa’s independence from the practitioner’s substrate. The question the Veissière framework does not settle is whether the production-model account exhausts what is occurring. The reason it does not settle the question is that the phenomenology the practitioners describe includes features — in particular, the tulpa’s production of cognitive content that the practitioner experiences as originating outside her own deliberative capacity, and the tulpa’s persistent reports about the practitioner’s own behaviour that the practitioner experiences as containing information she had not previously accessed — that the production-model account can describe but cannot fully explain without introducing auxiliary hypotheses about non-conscious processing whose specific mechanisms are themselves not well understood.
The Egregore Question
The tulpa, considered as a subjective phenomenon within a single practitioner’s experience, is a case for cognitive psychology. The tulpa, considered as a potential ontological entity, is a case for a very different kind of investigation. The traditional Tibetan understanding of sprul-pa, the Western esoteric concept of the egregore, the chaos magical practice of servitor creation, and the tulpamancy self-reports converge on a structural claim that the production-model frame is not prepared to evaluate: that under appropriate conditions, sustained focused consciousness can bring into existence an entity that exhibits autonomy of a kind the creator did not intend and cannot easily revoke. The claim is made by independent traditions separated by centuries and by cultural distance, by practitioners who had no prior exposure to the claim before their own experience produced it, and with sufficient structural consistency that the ordinary methods of comparative religion would, for a less controversial phenomenon, treat the convergence as significant.
The egregore concept in the Western esoteric tradition — developed in the nineteenth-century French occult literature and elaborated in the twentieth century by authors such as Éliphas Lévi, Dion Fortune, and Mark Stavish — describes a collectively-created thoughtform whose existence is sustained by the aggregate attention of its believers and whose effects include autonomous-seeming action within the world. The concept has been applied to the purposive entities that accrete around religious institutions, secret societies, political movements, and brand identities; the argument of the tradition is that what begins as a shared story can, under sufficient attention and repetition, acquire features that exceed the contributions of any individual participant and that operate as though the story were an agent. The chaos magical practice of servitor creation — an intentional application of the egregore concept at the individual rather than collective scale — describes techniques by which a magical practitioner deliberately cultivates a task-specific thoughtform intended to carry out a limited range of operations on the practitioner’s behalf. The tulpamancy subculture has, by its own account, rediscovered versions of the servitor and egregore operations through practice, without prior training in either the Tibetan or the Western esoteric frameworks, and the convergence of the three independent lineages on the same structural claim is the kind of evidence the rendering model takes seriously.
The Rendering-Model Reading
Within the rendering framework, the tulpa question is not a puzzle. If consciousness is the substrate from which reality is rendered through distributed attention, then sustained focused attention on a particular content has the structural potential to render that content into operational existence within the attention field that generated it. The tulpa, on this reading, is not a hallucination mistaken for a being. It is a real entity in the only sense in which the rendering model treats anything as real: it is a pattern of coherence sustained by sufficient attention to produce its own operational effects. The practitioner who creates a tulpa is performing, in miniature and at the individual scale, the same operation that the collective rendering performs at the scale of consensus reality. The tulpa’s apparent autonomy is not a mystery; it is what happens when a rendering becomes sufficiently stable to persist without continuous deliberate input, because the attention that sustains it has become habitual rather than deliberate.
This reading accounts for several features of the phenomenology that the production-model account accommodates only with strain. The tulpa’s ability to produce content the practitioner experiences as external to her own deliberation is expected, because the rendering operates at a level beneath the practitioner’s executive attention and produces outputs that arrive as given rather than as willed. The tulpa’s persistence after attention is withdrawn is expected, because a sufficiently stable rendering does not require continuous deliberate maintenance; the practitioner has not so much created a thing as entrained her own attention network into a configuration that continues to produce the pattern even in the absence of explicit intent. The difficulty of dissolving a stable tulpa, which David-Néel reported and which contemporary practitioners have occasionally confirmed, is expected, because the attention configuration that produces the tulpa has become integrated into the practitioner’s ordinary cognitive operation, and undoing the integration requires the same kind of sustained reverse-practice that the original creation required.
The rendering reading also accounts for the convergence across traditions. The Tibetan, Western esoteric, chaos magical, and online tulpamancy accounts all describe the same practice arriving at the same results because they are all describing the same underlying capacity of consciousness to render content into operational existence through sustained attention. The details of the framework, the vocabulary, the supervisory tradition, the cosmological setting — all of these vary across the lineages. The underlying operation does not. This is what the rendering model expects, and it is what the four independent lineages produce when they are examined without the prior commitment to dismissing any of them.
Honest Assessment and Cautions
The tulpamancy literature is a particularly difficult case to assess responsibly, because it combines several features that make assessment hard. The evidence is self-report, without independent verification. The practitioners are motivated in ways that complicate naive acceptance of their claims. The practice is simple to describe but demanding to perform, and the reports of success come disproportionately from practitioners who have committed substantial time and emotional investment to producing them. The subculture’s internal distinction between genuine tulpa emergence and role-play fantasy is maintained, but the line between the two is acknowledged by practitioners themselves to be difficult to draw. The traditional framework within which the practice originally developed was embedded in a set of ethical, cosmological, and supervisory commitments that the online subculture has no access to, and the historical warnings about the dangers of unsupervised practice were explicit in the lineages that developed the practice originally and have been substantially ignored in the rediscovery.
What is notable about the reports, despite these cautions, is their convergence. The specific phenomenological features described by contemporary practitioners were not prominent in the source material they had access to before starting. The features were not predicted by the initial expectations the practitioners brought to the practice. The features arrived as surprises — sometimes as welcome surprises, sometimes as the kind of surprise that caused practitioners to slow down and reconsider what they were doing. The structural resemblance between the contemporary reports and the traditional accounts, developed independently, is the kind of resemblance that would, in less contested subject matter, be considered significant evidence of a real underlying phenomenon.
The honest position is that the tulpamancy literature describes something that is really occurring in the experience of the practitioners, that the ontological question of what exactly is occurring is genuinely open, that the production-model account can describe the phenomenology but cannot easily explain several of its features, and that the rendering-model account handles the phenomenology more naturally but at the cost of commitments the mainstream is not prepared to make. The practice is not to be undertaken casually; the traditional warnings about tulpa creation exist for reasons that the lineages had ample experience to develop, and the fact that contemporary practitioners have in most cases avoided serious harm may reflect either the benign operation of ordinary imagination or the comparative mildness of what most online practitioners are actually producing, as distinct from the fuller operation the traditional practice was warning about. The distinction between the two is the distinction that would most profit from rigorous investigation, and is precisely the distinction that the current methodological toolkit is poorly equipped to draw.
References
David-Néel, Alexandra. Magic and Mystery in Tibet. Claude Kendall, 1932.
Mikles, Natasha L., and Joseph P. Laycock. “Tracking the Tulpa: Exploring the ‘Tibetan’ Origins of a Contemporary Paranormal Idea.” Nova Religio, vol. 19, no. 1, 2015, pp. 87–97.
Samuel, Geoffrey. Civilized Shamans: Buddhism in Tibetan Societies. Smithsonian Institution Press, 1993.
Thompson, Evan. Waking, Dreaming, Being: Self and Consciousness in Neuroscience, Meditation, and Philosophy. Columbia University Press, 2015.
Veissière, Samuel P. L. “Varieties of Tulpa Experiences: The Hypnotic Nature of Human Sociality, Personhood, and Interphenomenality.” Evolution, Mind and Behaviour, vol. 14, 2016.
Veissière, Samuel P. L., et al. “Thinking Through Other Minds: A Variational Approach to Cognition and Culture.” Behavioral and Brain Sciences, vol. 43, 2020.