Is NLP evidence-based? What the research actually shows — and where NLP's power comes from.

In 2002 I led an Indian team to the Austrian Youth Red Cross's 47th International Study and Friendship Camp at Langenlois. The following year I began teaching in Switzerland, where I would return for six consecutive years. I had come from a village in the Himalayan foothills of Jammu, and I stood in front of European rooms without robes, without a spiritual title, without the guru costume Indian teachers abroad were expected to wear. Just the work. In September 2004, NLP-Akademie Schweiz named me a co-operating Partner and Sponsor in research and training, and I have been teaching on two continents since.

In those rooms, over more than two decades, I kept meeting the same puzzle. Two practitioners hold the same certificate. Both learned the same anchoring procedure, the same reframe, the same language patterns, often from the same manual. One of them changes the person sitting opposite. The other performs a ritual and nothing moves.

The difference is never the technique. Both people have the technique.

I have been studying NLP since 1990, and my answer is the argument of this essay: a technique carries only as much power as the tradition it was drawn from, and only when the practitioner has touched that tradition directly. NLP's founders did not invent its methods out of air. They modelled them — from a hypnotist in Phoenix, from two therapists in California, and at one remove from contemplative systems that had been refining the same moves for a thousand years. Train only in the extracted technique and you inherit the pipe. Train in what it was extracted from and you inherit the well.

What NLP is, in one honest paragraph

Neuro-Linguistic Programming is the study of the structure of subjective experience: how a human being codes perception, language and behaviour, and how that coding can be changed. It was developed in the mid-1970s at the University of California, Santa Cruz, by Dr. Richard Bandler and John Grinder, who built it by modelling three practitioners whose results outran the theory of the day — Fritz Perls, Virginia Satir and Milton H. Erickson, the most influential clinical hypnotist of the twentieth century. The first fruits appeared as The Structure of Magic and Patterns of the Hypnotic Techniques of Milton H. Erickson, both in 1975.

Notice what that history tells you. NLP began as an act of extraction. Bandler and Grinder watched masters, found the structure inside the artistry, and wrote it down so it could be taught. That makes NLP unusual among change disciplines: it is a modelling discipline, a method for finding out how excellence works wherever excellence is found.

Which raises the question the field rarely asks itself. If NLP's power came from modelling deep sources — what happens when you take it back to sources deeper still?

Three kinds of authority, kept apart

The evidence, stated plainly

Start with the part that does not favour me, because you will find it anyway.

Direct trial evidence for NLP-branded protocols is thin. A 2012 systematic review by Sturt and colleagues in the British Journal of General Practice searched for health-outcome studies of NLP techniques and found little of sufficient quality to draw conclusions from. I would rather you heard that from me than found it on Wikipedia after paying someone.

If a trainer tells you research has proven every technique they sell, they have failed the test in front of you.

Now the part that is rarely put beside it. The disciplines NLP was modelled from test very well.

The parent discipline first. A 2024 umbrella review by Rosendahl and colleagues gathered forty-nine meta-analyses covering 261 randomised controlled trials of hypnosis, with the strongest support in pain and procedural distress. In the United Kingdom, NICE clinical guideline CG61 lists hypnotherapy among the options for irritable bowel syndrome that has not responded to first-line care — a mainstream health authority naming trance work in an official pathway.

The clinical well hardly needs defending. Cognitive behavioural therapy sits in first-line positions across NICE guidance for depression and the anxiety disorders. For post-traumatic stress, the World Health Organization's 2013 guidelines and the American Psychological Association's 2017 guideline both name trauma-focused therapies, EMDR among them.

The contemplative practices have crossed into this column too. Kuyken and colleagues' 2016 meta-analysis in JAMA Psychiatry, pooling individual patient data from nine trials and more than 1,200 people, found that mindfulness-based cognitive therapy reduced the risk of depressive relapse — which is why NICE recommends it for exactly that purpose. Lazar and colleagues (2005) found thicker cortex in attention-related regions in long-term meditators. Hölzel and colleagues (2011) measured increased grey-matter density after a standard eight-week course. Lutz and colleagues (2004) recorded high-amplitude gamma synchrony in long-term Tibetan practitioners that novices did not show.

And the physiology. In 1982, Herbert Benson's team reported in Nature that g Tum-mo practitioners in Upper Dharamsala raised the temperature of their fingers and toes by as much as 8.3 °C at will. Read that precisely: peripheral temperature, attributed to vasodilation. It is widely misreported as raising body temperature, and it did not — core temperature was unchanged, a distortion a 2013 PLOS ONE paper documents specifically. The accurate version is the more interesting one: voluntary control of peripheral circulation, measured under observation, published in one of the two most demanding journals in science.

So both things are true. The branded protocols are under-tested; the sources are well tested. That is not a contradiction once you remember what kind of discipline NLP is. A modelling discipline is tested by testing what it models. The wells hold up under instruments. What varies wildly is the practitioner.

Which is why the question that matters is never "does NLP work?" It is "has this trainer drunk from the wells, or only handled the pipes?"

Six wells, one discipline

Strip any NLP certification to its working parts and you find six dimensions. Trance, the states in which change happens. Language, the patterns that direct attention and meaning. Physiology, the body as the fastest lever on state. Psychology, the clinical understanding of how minds get stuck and come unstuck. Organisational application, how it scales from one chair to a boardroom. And modelling, the master skill of extracting structure from excellence.

Here is the position I have built my school on, and some colleagues will dispute it: you cannot teach any of those six at full strength if NLP is the only place you ever studied it. Much NLP training is a photocopy of a photocopy — a weekend certificate teaching the names of techniques whose sources the trainer has never touched. The labels survive the copying. The power does not.

So I did it the long way. Trance through board-certified clinical hypnotherapy with the National Guild of Hypnotists, where I hold instructor certification and founded its India chapter. Language and modelling at the source, training to Practitioner, Master Practitioner and Trainer under Dr. Richard Bandler and John La Valle. Psychology formally — a master's in organisational behaviour, postgraduate clinical training, and later a doctorate. Physiology and state for fourteen years inside an Indian contemplative lineage under Dr. Ajay Kotwal, alongside yoga certification from the Sivananda school and Tai Chi and Qigong in the line of Bruce Kumar Frantzis.

What follows are those wells as I know them from the inside, and for each one the places where its water is already flowing through the NLP you were taught, usually without anyone saying where it came from.

Well I · Clinical hypnosis: the acknowledged parent

Begin with the well NLP openly admits to. In 1974 Bandler and Grinder went to Phoenix to study Milton H. Erickson, and published the patterns the following year. What the field calls the Milton Model is not an NLP invention. It is Erickson's speech, transcribed, dissected and arranged so that ordinary mortals could learn it.

Every pattern in it is a hypnotic instrument. Pacing and leading is the spine of Erickson's conversational inductions. Embedded commands hide a directive inside an innocent sentence. Presuppositions build the outcome into the grammar, so arguing with the sentence means accepting its premise. The double bind offers two roads to one destination. Analogic marking lets the voice underline words the conscious mind never notices being underlined.

Here is what a certificate in those patterns does not give you. It does not give you a formal induction — the eye-fixation method Braid documented in 1841, or Dave Elman's rapid induction, taught to physicians and dentists for exactly that reason. It does not give you fractionation, the observation that bringing a person out of trance and back in takes them deeper each time. It does not give you ratification: the arm catalepsy, the time distortion, the small demonstrations that convince a sceptical nervous system something real is happening. And it does not give you the depth scale, down to the state in which nineteenth-century surgeons operated without anaesthetic.

Once you have those, the toolbox reorganises in your hands. Future pacing turns out to be post-hypnotic suggestion under a newer name. An anchor set during genuine trance behaves differently from one set during polite eye-closure: it holds.

I will say something that will annoy part of my own profession. A practitioner who has never taken a client through a formal hypnotic induction has no business teaching the Milton Model. They are reciting poetry in a language they cannot speak. I trained as a Board Certified Hypnotherapist and then as a Certified Instructor with the National Guild of Hypnotists — founded in 1951, serving over ten thousand members in 105 countries — and founded its India chapter so Indian practitioners could reach the parent discipline rather than the extract.

Well II · Clinical psychology: the sibling raised in the laboratory

Two of NLP's three original models were clinicians. NLP was born wearing the clinic's coat and then — this is my reading, and you may dispute it — walked out of the clinic and mostly stopped talking to it. That was a loss for both sides.

Sit with the clinical literature for a few years and the structural rhymes are everywhere. The Meta Model's challenges — who specifically? compared to what? how do you know? — rhyme with Socratic questioning in cognitive behavioural therapy and with the downward-arrow technique that chases a thought to the belief beneath it. Parts work rhymes with the empty-chair dialogues Perls ran, which is direct parentage, and with the mode work of schema therapy. The well-formed outcome rhymes with the miracle question of solution-focused brief therapy. The ecology check is a behavioural experiment run in imagination before it is run in the world.

Let me apply my own rule before anyone applies it for me: a rhyme is not a validation. The trials behind CBT tested CBT, not the Meta Model. What the rhymes tell you is something else and still valuable — that when Bandler and Grinder extracted structure from gifted clinicians, they extracted things the clinical world would independently converge on and then spend decades testing.

What formal psychology adds is mostly what a weekend certificate cannot: the discipline of restraint. Assessment before intervention. Knowing the difference between a limiting belief and a clinical condition that belongs with a psychiatrist. Structured protocols with a beginning, middle and end. Supervision. Contraindication. The professional habit of asking "should I?" before "can I?"

My own route ran through a master's in organisational behaviour, postgraduate clinical training, registration with India's Rehabilitation Council as a Clinical Psychologist (Assoc.), CRR No. A100310, and a PhD in Buddhist Studies. None of it was required to hold an NLP certificate. All of it changed what happens when I sit down with a human being in difficulty.

The Council registers the practitioner, not the method. It does not endorse Neuro-Linguistic Programming, and nothing here should be read as suggesting otherwise.

Well III · Yoga: state before content

NLP defines state as the product of internal representations and physiology, then spends most of its training hours on the representations. Yoga spent roughly two thousand years on the physiology and left an engineering manual.

Consider yoga nidrā. The practitioner lies still while attention is rotated through sixty-one points of the body; the body falls toward sleep while awareness stays lit; and into that state a saṅkalpa is placed — a short, positively worded intention, delivered at the moment of maximum receptivity. Any hypnotherapist recognises the architecture instantly: induction, deepening, suggestion. The yogis built it independently and understood something many hypnosis manuals still under-teach: the state must be established before the content is delivered, and the depth of the state decides the fate of the suggestion.

Or consider breath. Prāṇāyāma is a taxonomy of state changes keyed to the nostrils and the diaphragm — nāḍī śodhana to balance, bhastrikā to arouse, bhrāmarī to settle, kumbhaka to still the whole system. Breath is the one lever of physiology that is both automatic and voluntary, which makes it the fastest handle on state a practitioner will ever hold.

Attention is the third gift. Every trainer meets students who say "I can't visualise." Almost none of them lack the capacity; what they lack is trained attention, and yoga treats attention as a muscle with a progressive overload programme — trāṭaka, the unblinking gaze; dhāraṇā, attention fixed until the point dissolves. A few weeks of it and the same student's images arrive bright and stable. The problem was never the cinema. It was the projectionist's hands.

And anchoring has grandparents here. Nyāsa places mantra on the body, point by point, until touch itself carries state. Mudrā seals a state into a gesture. Kinaesthetic anchors with a thousand years of field testing, set within a system that also supplies what NLP calls ecology: preparation before technique, ethics before power. I hold the title Yoga Siromani from the International Sivananda Yoga Vedanta Centre, conferred at Neyyar Dam in December 2010 — awarded for services rendered to the propagation of yoga rather than for completing a course.

Well IV · Vedānta: the map, the territory and the rope

Every NLP training on earth opens with Alfred Korzybski's line from Science and Sanity: the map is not the territory. A fine sentence. Śaṅkara's tradition had been teaching the same epistemology for over a thousand years with a better prop — a rope on a dusk-lit path, mistaken for a snake. The fear is real. The sweat is real. The snake never existed. Advaita Vedānta even names the full manoeuvre: adhyāropa-apavāda, deliberate superimposition followed by deliberate negation. First understand how the false map was laid over the territory; then peel it off. That is the logic underneath every reframe you will ever run.

Vedānta also anticipated the Meta Model at its most serious point of application: identity. When a client says "I am a failure," a practitioner hears a nominalisation welded to a complex equivalence and starts asking recovery questions. The Upaniṣadic method of neti neti is the same challenge aimed at the largest identity statement a mind can make. The Meta Model interrogates a sentence. Neti neti interrogates the speaker.

Then the witness. NLP teaches dissociation and third position — useful, and rented by the minute. Vedānta's sākṣī-bhāvanā trains the witness as a permanent residence: not a technique you visit under instruction but a stance that stays put at two in the morning when no practitioner is in the room. One is a state. The other is development. Confusing the two is, in my view, the most common error in the personal-change industry.

And Vedānta explains something every trainer has watched with a sinking heart: why weekend learning evaporates. Its syllabus for knowledge that actually restructures a person has three stages — śravaṇa, listening until the idea is accurately received; manana, interrogating it until your own objections are exhausted; nididhyāsana, dwelling in it until it stops being information and becomes the place you think from. Most certification culture stops at stage one. The fade-out that follows is not a mystery. It is a curriculum design error, diagnosed centuries in advance.

Well V · Kashmir Śaivism: installation, not visualisation

This well rises a few hundred kilometres from the village where I grew up. In ninth-century Kashmir, Vasugupta set down the Śiva Sūtras; a century later Abhinavagupta composed the Tantrāloka, still one of the most complete manuals of consciousness engineering ever written. Where Vedānta reaches its goal by negation, the Śaiva tradition reaches it by construction — and construction is exactly the business NLP is in.

Its central instrument is bhāvanā: contemplative imagination, defined with a rigour that should embarrass most modern visualisation coaching. Bhāvanā is not picturing a desired outcome. It is dwelling in a recognition as already real, repeatedly, until identity reorganises around it. Set that beside NLP's "acting as if," its future-self work, its new behaviour generator, and you find the same move with one difference: the Śaiva texts treat it as installation of identity and engineer it accordingly — preparation of the body, preparation of the state, and only then the installation. Technique last. Everything NLP calls "setting the frame," this tradition calls consecration, and it refuses to skip it.

The tradition also compiled what I can only call a state-change library. The Vijñānabhairava lists one hundred and twelve doorways into expanded awareness — attention in the gap between two breaths; the moment a sound dies into silence; the pause between one thought ending and the next arising; the instant of a sudden fright, caught before the story starts. NLP discovered, correctly, that an interrupted pattern opens a door. This text catalogued a hundred and twelve doors, roughly twelve centuries earlier.

One more Śaiva gift, to my mind the most practical: the doctrine of the four upāyas, four means of transformation matched to four kinds of student. For one, awareness alone suffices. Another needs contemplation of energy. A third needs effortful technique — breath, posture, mantra, repetition. A fourth needs nothing but the teacher's presence. Twelve hundred years before anyone said "pace the client's model of the world," this tradition had already ruled that teaching one method to every student is malpractice.

Well VI · Tibetan Buddhism: modelling taken to its limit

My doctorate is in Buddhist Studies, and my thesis examined enlightenment in Vajrayāna Buddhism, so I will confess a bias upfront: I think Tibet built the most complete change technology on this list. Judge for yourself.

Start with modelling, NLP's crown jewel. In deity yoga's generation stage, the practitioner constructs an exemplar of awakened capacity in complete sensory detail — form, colour, sound, seed-syllable, environment — then steps into it and lives from it, session after session, until its qualities begin answering to their name. NLP's deep trance identification is this manoeuvre with the iconography removed. But Tibet adds the step NLP forgot: the completion stage, in which the constructed identity is deliberately dissolved back into open awareness. Build the model, wear the model, and then — this is the mastery — take the model off. A discipline that only installs and never dissolves produces impressive people who are stuck being impressive. The Tibetans engineered against that failure mode a thousand years ago.

Perceptual positions have a Tibetan ancestor too, with higher stakes. Tonglen has the practitioner breathe in another person's suffering and breathe out relief toward them — second position practised not as an information-gathering trick but as a direct assault on self-fixation, sixty seconds at a time. And reframing has one: lojong, the mind-training tradition codified in twelfth-century Tibet as fifty-nine slogans, which amounts to a deck of field-tested reframes, memorised and deployed at the exact moment life provides the trigger.

Underneath all of it sits attention, treated as a trainable skill with a written syllabus: the nine stages of śamatha, from "placing the mind" to effortless equanimity, each stage with its named faults and named antidotes. Every visualisation-dependent technique in NLP stands on the quality of the student's attention, and Tibet is the tradition that wrote the progression manual.

How to choose an NLP trainer: the depth test

So you are considering a Practitioner or Master Practitioner certification, in India or anywhere. Five questions will tell you nearly everything. I invite you to aim all five at me.

One. Who signs the certificate? Ask whether the trainer's own credentials trace to The Society of Neuro-Linguistic Programming™, the licensing body of the co-creator. Mine carry the signatures of Dr. Richard Bandler, Kathleen S. La Valle and John J. La Valle, with seal numbers on the documents. I am not the only trainer in India who can say that, which is exactly why you should ask — because many cannot.

Two. Has the trainer done formal hypnosis? Real inductions, real depth work, real clients — the parent discipline, not two Milton Model afternoons. Ask how they were certified and by whom.

Three. Is there clinical grounding? Formal psychology changes what a practitioner notices, and more importantly teaches them when the person in front of them needs a psychiatrist rather than a technique. Ask what their plan is for the client they should not take.

Four. Is there contemplative depth with a named lineage? Not a retreat photograph. A teacher's name, a tradition, and years — the kind of answer that can be checked. Mine is fourteen years under Dr. Ajay Kotwal, alongside Sivananda yoga and a doctorate in Buddhist Studies, and every item sits documented on the credentials page.

Five. Watch how they handle evidence. A trainer who tells you research has proven everything they sell has just failed in front of you. The honest answer is layered, exactly as this essay laid it out. Depth and honesty travel together; I have never met one without the other in this profession.

Filed under practitioner experience, weighed as testimony. This training was built in Jammu, with village students, long before it had a website. One of those students went on to win MTV Roadies Real Heroes in 2019 and has said on his own filmed record that the work changed his life. The Stress & Happiness Management Workshop I have run in Switzerland has drawn participants from more than a hundred countries over the years it has run. The pattern has held everywhere I have taken it: source-depth in the trainer becomes speed in the student.

Thirty-six years ago I lowered the first bucket. Since then the route has run through Bandler's trainings, the Guild's clinics, the clinical registers, the Sivananda mats, the Śaiva texts of my own hills and the Tibetan syllabus of attention — six wells, one discipline. The pipes are printed in every manual on earth.

The water is not in the manual. It never was.