Are you enough of a sceptic to be sceptical about your own scepticism?

Andy Bradbury

Andy Bradbury is a social psychologist by training and has spent most of his working life in personnel and/or training.
He is the author of 'Develop Your NLP Skills', Kogan Page, London: 2010 (4th edition)


(This is a lightly edited version of an article which appeared in Dr. Michael Heap's magazine The Skeptical Intelligencer, Vol 11, December 2008. pp.14-26.  For example, some quotations, duplicated from the main text, which I felt interferred with the flow of the discussion have been removed; and some links to some clarifying material elsewhere on the website have been added.  Alterations are indicated by the use of "maroon-coloured" text.)


I really found it quite hard to write this article.  For two reasons.
Firstly, throughout my exchange of e-mails with Dr Heap he has been unfailingly courteous and friendly.  This included allowing me to see the initial version of his own article* back in the early autumn so that I could take account of it here.
[*   Heap, M. The validity of some early claims of Neuro-Linguistic Programming, The Skeptical Intelligencer, Vol 11, December 2008. pp.6-13]

I take no pleasure, therefore, in stating that virtually everything he, in common with several other oft quoted critics, has written about the field of NLP is in error.  Moreover, since Dr Heap has framed this discussion in terms of the accuracy of his judgement as to the validity and implications of the experiments he reviewed, I have no option, if I am to answer him adequately, but to call that judgement into question.
Secondly, precisely because there are so many errors, I hardly know where to start to set the record straight.


The comments I make here will, as far as possible, be based on NLP as presented by its co-creators, Richard Bandler and John Grinder - including a number of points I was able to discuss in person with John Grinder in September 2008 - rather than NLP according to Andrew Bradbury or any other third party.  For reasons explained below I will ignore all references to writers such as Stephen Lankton, John Seymour, Joseph O'Connor and Steve Andreas, cited by Dr. Heap.

Evaluating the claims

When I first communicated with Dr Heap it was specifically in regard to the two papers he wrote in the 1980s, 'Neurolinguistic programming: An interim verdict' (1988), and 'Neurolinguistic programming: What is the evidence? (a paper delivered in 1987 but not published until 1989).  Since both the 'Interim verdict', and the new paper, 'The validity of some early claims of neuro-linguistic programming' (2008), are effectively variations on the original, for the sake of brevity I will refer to them here, in order of publication, as IV1, IV2 and IV3.

For the benefit of anyone not familiar with IV1, this is the only version which contains a full list of references to the material (all Abstracts) originally reviewed, and is available (along with IV2) on Dr Heap's web site.  The Abstracts in question can nearly all be found in this research database.

In reverse order

In his latest paper, Dr Heap has provided an extended introduction to the main points of IV1 and IV2, most of which is either inaccurate or irrelevant.  It is useful in the current discussion, however, in that it gives further clues as to how Dr Heap and other critics of 'NLP' may have arrived at their erroneous conclusions.

The assertions about the current state of the field of NLP, for example, have the appearance of 'truth' though in fact they show a substantial mismatch with the facts. Thus:

1. '... in one of my papers (Heap, 1989, pp. 118-119) I gave the following description of NLP, one that I still consider accurately portrays how NLP at that time represented itself:

"[NLP] is a model of human behaviour and cognition ..."'
(Heap, 2008. p.7)

In the first place NLP is a specific modelling process, not a person or an organization, and therefore has never 'represented itself' as anything at all.  I make this point, obvious as it may seem, because critics often use the notion of NLP as a 'thing' as a means to justify quoting almost any source going, just as long as it mentions NLP, as though all sources were equally representative and accurate.  In fact Dr Heap himself seems to have taken this position (see below).  In practice, however, this premise is simply not sustainable.  I have personally read and reviewed over 150 books on the field of NLP for my own web site - and the variations on a theme vary from slight modifications to complete revisions.  By 'cherry picking' one's quotes one could probably show that these nebulous 'NLP writers' claim that black is white and day is night, so to speak.  But such an approach would bury the details of authentic NLP, techniques and applications under a mountain of dross.

And in the second place, NLP has in fact always been a specific modelling process or technique.  It is not, itself, a model of anything; though I'm happy to accept that, based on his own mental maps, Dr Heap honestly believed/believes that his definition is accurate.

2. 'There is no mention of [Bandler and Grinder's works] in learned textbooks or journals devoted to these disciplines.' (Heap, 2008. p.6)

This claim is plainly untrue, since several of the references in IV1 and IV2 were from such sources (i.e. specialist journals).  But there is an even more basic reason - different writers often use different labels and/or don't know the original source of their material!  For example, in the first section of an article entitled 'How to get exactly what you want' in a recent issue of New Scientist (May 10, 2008) we were introduced to something called 'mimicry' as a persuasion technique.  The article included references to the Journal of Experimental Social Psychology, vol 44, p 461 and to the Journal of Consumer Research, vol 34, p 754.

However, no mention was made of NLP or Bandler and Grinder by name, so people with little or no knowledge of NLP would be unlikely to recognise that 'mimicry' - which was described in the article as being the subject 'of recent studies' - is nothing more than two or three of the techniques that are referred to in NLP jargon as 'pacing' - a group of techniques that began to be used by 'NLPers' more than 25 years ago. So close is the match between 'mimicry' and some aspects of 'pacing' that certain instructions in the New Scientist article were almost word for word what I was taught about pacing on an NLP-oriented business training course in the late 1980s.

As a matter of interest, even as I was working on this article I was made aware, on one of the NLP chat groups, of an even more recent unequivocally NLP-related article in a professional journal: Using a modified neurolinguistic programming swish pattern with couple parasuicide and suicide survivors, Gerald A. Juhnke, Kenneth M. Coll, Michael F. Sunich, and Ronda R. Kent.  The Family Journal, Oct 2008; vol. 16: pp. 391-396.
For more material of a similar nature, a list of serious NLP-related research papers collected by Dr. Paul Tosey of the University of Surrey can be found here: , and a list of relevant education-related articles are included at the end of FAQ 22.

3. An even more basic error is the subject of the next claim:

'Neither is any of this material taught on psychology courses at pre-degree and degree level.'
(Heap, 2008. p.10)

Quite apart from the students who e-mail me to ask questions about NLP that have come up on their degree courses, there are a growing number of places where NLP is now a degree course subject in its own right.  NLP gets plenty of attention at the University of Surrey, for example, and at least two people at that institution have taken investigations of NLP as their Ph.D. projects. (Both candidates came up with results which supported the validity of the techniques being investigated).  NLP also features in courses at the University of Portsmouth, the Open University and the University of Kingston in the UK.  In Australia there is a post-graduate course in NLP run jointly by the NLP training company, Inspiritive, and the University of Sydney, and those two organizations are also working together on 'scientific' studies of various NLP-related techniques and concepts.
[I know that that there are also universities/colleges in the US which run NLP-related degree courses, but I do not have the details at this time.]

    (I have the word 'scientific' in inverted commas because I do not believe that we yet have investigative techniques suitable for testing 'pure' psychology. See Gazzaniga (1998, Preface), Frith (2007, Prologue) and my FAQ on the 'eye accessing cues' at for further discussion of this claim.)

4. Dr Heap's article also, somewhat ingenuously, I think, asks, 'Why this almost total neglect of a body of knowledge that, if it has any authenticity, should occupy a pivotal role in the study of human psychology?'
(Heap, 2008. p.10)

    One possible reason for many people's attitude towards NLP today is that it was 'trashed,' in the 1980s and '90s, by people who supposedly knew what they were talking about.  It seems that many readers - lay people, academics and scientists alike - don't always take the time to check the accuracy of material when it is presented by people who are judged to be authoritative/reliable within their own field of study.
Let me give an example:

    In IV1, Dr Heap returns several times to studies which dealt with the "stability" of subjects' PRS (see comments on Birholtz (1981) and Ridings (1986), on page 270, for instance), despite the fact that Bandler and Grinder had specifically indicated that a person's PRS is context-related and is therefore liable to change:

In a particular context you will be aware of* one [sensory] system more than anpther.  I assume that when you play athletics or make love, you have a lot more kinesthetic sensitivity.  When you are reading or watching a movie, you have a lot of visual consciousness.  You can shift from one to another."
(Frogs into Princes, 1979. p.36. Italics as in the original text.)

    *It is, in fact, the representational system(s) one is aware of which constitute the PRS, of course (see, for example, Frogs into Princes (1979), page 28.)  And time itself can be an important factor in determining the difference between two contexts even when they might otherwise appear to be more or less the same.  Yet Dr Heap, despite having done "some basic training" and having "read some of the main texts at that time" (which included Frogs into Princes - first sentence of IV3, page 6 - says nothing about the experimenters' failure to notice that they were totally ignoring, or simply weren't aware of, the claims that Bandler and Grinder had made on this subject.  Indeed, did Dr Heap himself notice the blatant contradiction?

    This really goes to the heart of the question of whether people like Dr Heap, Professor Levelt (of the Max Plank Institute in Nijmegen, Holland) and others of similar standing, are really qualified to make objective judgements about NLP.  I do not for one moment question the fact that these two men are genuine experts in their own fields of study - hypnosis and psychotherapy for Dr Heap, and psycholinguistics for Prof. Levelt.  But how does that qualify them as experts on NLP?  In two words, 'It doesn't'.  In fact both men have proved themselves notably short on expertise when it comes to NLP, possibly because (a) it is outside of their specialist areas of knowledge, and (b) it seems that neither of them did what was required to bring themselves up to speed on the subject.

    Despite its substantial inaccuracies, Dr Heap's review has nevertheless been quoted innumerable times as though its content was reliable, for example by hypnotherapist Dr. (as in D.Phil. rather than M.D.) Dylan Morgan, writing for the Journal of the National Council for Psychotherapy and Hypnotherapy Register, Spring issue of 1993 (still accessible on his web site), and by Dr. Robert Carroll, who uncritically cited Morgan's commentary on Dr Heap's allegedly authoritative review in the NLP article on his so-called Skeptic's Dictionary web site (repeated in his subsequent book of the same name).  And of course all three items have been referenced on Wikipedia at one time or another on the NLP-related pages.
    This is indeed 'the blind leading the blind.'

    A second factor may well have been that neither Bandler nor Grinder had the least interest in joining 'the establishment', so to speak.  In fact John Grinder specifically told me, during our recent conversation, that he and Bandler presented their first two books - The Structure of Magic, Vols 1 & 2 - in an academic style, just to show that they could do it, and thereafter carried out their studies with a complete lack of interest in any academic opinions regarding their work.

    Given the typical response of academics to this kind of attitude - namely 'our way or the highway' - a significant mismatch between the majority of academics and the developers of NLP was pretty inevitable.  For an example of this clash of attitudes in action see the comments in Dr Heap's current paper such as:

'To arrive at these kinds of generalisations about the human mind and behaviour would certainly require prolonged, systematic and meticulous investigation of human subjects using robust procedures, observing, recording and analysing the phenomena under investigation.  There is just no other way of doing this.  Yet, when they made their assertions, the authors never revealed any of this to their students and to their readers; they merely stated that this was what they had noticed.'
(Heap, 2008. pp.6-13.  Italics added for emphasis)

    In the first place, this is a highly idealised account of how research works.  And in any case, compare this statement with Bandler and Grinder's own take on the subject in Frogs into Princes (remembering that the book is based on transcripts of actual NLP-related training seminars, edited by Steve Andreas, and therefore gives us fairly direct information about what Bandler and Grinder were saying to their students in 1978):

'You ask somebody a question.  They say 'Hm, let's see,'and they look up and to their left, and tilt their head in the same direction.  When people look up, they are making pictures internally.
'Do you believe that?  It's a lie, you know.  Everything we're going to tell you is a lie.  All generalizations are a lie.  Since we have no claim on truth we will be lying to you consistently throughout this seminar. ...
'As modelers, we're not interested in whether what we offer you is true or not, whether it's accurate or whether it can be neurologically proven to be accurate, an actual representation of the world.  We're only interested in what works.'
(p. 18)

    In other words, the criticism is not justified because it addresses an entirely different kind of claim than the one Bandler and Grinder were making.  The attitude that: 'you must do research the one and only way we regard as valid' might well be okay - IF someone is claiming that they are dealing in 'proven facts', 'the truth' or whatever.  But Bandler and Grinder had not made such claims.  They were indeed deliberately offering their generalisations as nothing more concrete than 'what we have noticed.'  Far from hiding anything from their audience, as Dr Heap seems to be implying, Bandler and Grinder have been consistently open and honest about their lack of interest in academic-style research.

5. In another part of the 2008 article we find a reference to 'Bandler & Grinder (1979, p 40):

'A lot of school children have problems learning simply because of a mismatch between the primary representational system of the teacher and that of the child.  If neither of them has the flexibility to adjust, no learning occurs.  Knowing what you know now about representational systems, you can understand how it is possible for a child to be 'educationally handicapped' one year, and do fine the next year. ...'
(quoted in Heap, 2008. p.11)

    Now, I understand that some people will indeed find this claim 'extraordinary', but that is a very relative word, and in my experience usually means nothing more than: 'This doesn't fit with the way things 'are' in my view of the world.'  Thus it is pertinent to ask, 'How well-informed is this point of view?'  Have the people making this kind of statement ever read the book Righting the Educational Conveyor Belt, by Michael Grinder (1986 [original version], 1989 [first full-size paperback version]), which explains in detail how NLP-related techniques can be applied in education?
(Michael Grinder is John Grinder's brother.)

    And on a very practical note, have they read about the 'Durham NLP Project 2006' when an NLP-oriented approach was tested 'in the field'.  A quite detailed report of this application of NLP-related techniques in schools can be found online here:

6. Like many other critics, Dr Heap seems less than impressed with the suggestion that NLP-related techniques can produce high-speed results (e.g. see Heap, 2008. p.7).  But have these people read Richard Bandler's book Magic in Action and/or viewed the accompanying videotape (which must be obtained separately).  What they would find is a record of three brief sessions (all under 30 minutes) in which Richard Bandler deals with three cases which have previously proved intransigent in the face of therapeutic intervention.  I offer this as evidence with some confidence, since all three sessions were conducted under laboratory conditions and monitored and videoed by faculty members at Marshall University in West Virginia, USA, and the results checked for durability approximately 8 months later.

    Note, although NLP is not a form of therapy, Bandler uses the NLP-related techniques as adjuncts to the therapeutic process during these sessions.

    I might also point out that there appears to be an ongoing link between Marshall and NLP, since a recent (September 2008) check on Marshall University's web site listed a certain William A. McDowell as Professor of Counseling, with a Trainer qualification in Neuro-Linguistic Programming, and describes him, amongst his various publications, as 'the producer/-developer of 20 studio videos in Neuro-Linguistic Programming ...'.

7. And lastly, in this section, I refer you to point 4 of Dr Heap's section, Further developments of NLP, where he writes:

'NLP training courses abound and NLP now seems to be most influential in management training, lifestyle coaching, and so on.  Particularly with reference to this, the term "growth industry" appears to be apposite.
'I know little about this last-mentioned area of work but I am intrigued by this gradual extension of NLP beyond psychotherapy.  This may have something to do with the fact that the supply side of the market for psychological therapies looks pretty much saturated and the major potential customer in the UK at least, namely the National Health Service, tends to favour a limited range of products, notably those that are labelled 'evidence based'.'
(Heap, 2008. p.12)

    This is, to put it mildly, a pretty definitive example of how critics of NLP tend to use their own 'mental maps' to power their criticisms, with little or no regard for what is happening in the 'real world'.
Despite his admission that he knows 'little about this last mentioned area of work', Dr Heap promptly offers a description of what is going on, and an equally fictitious 'explanation' for why it is happening.  But once again, as happens whenever he presents us with some aspect of his 'NLP = therapy' map, the explanation is hopelessly wide of the mark.

    The use of NLP and NLP-related techniques in business is my own particular area of investigation and practice, and I have written an introductory 'how to' guide to the subject (now called Develop Your NLP Skills) which was first published over ten years ago - January 1997, to be precise.  Several other books of a similar nature came out in the early 1990s, including O'Connor and Seymour's Training with NLP (1994) and Kerry L. Johnson's Selling with NLP (1994).  A third book published that same year was Robert Dilts' Effective Presentation Skills, of particular interest here since it is little more than a description of a number of NLP-related techniques and was based on the training materials Dilts used in a seminar previously 'designed for the Italian National Railways as part of its efforts to becoming [sic] a 'learning organization'' (p xi).

    But of course these came some years after Dr Heap's review of 1988, and therefore he couldn't have known about them at that time.  What he could have known about, however, was Genie Laborde's books on NLP in the workplace: Influencing with Integrity (1983), and Fine Tune Your Brain (1988).  And he could have known about all of these books - and several others of a similar nature, such as John Grinder and Michael McMasters' Precision (1994) - by the time he came to write his latest paper.  Thus obviating the entirely erroneous claims quoted above.
Clearly, then, in making these claims about the development and progress of the field of NLP, Dr Heap has once again demonstrated the thorough lack of knowledge upon which he has constructed his supposedly well-informed discussion of those topics.
And, unfortunately, there's more.

The heart of the matter

    Before answering Dr Heap's 1988 claims, it is necessary that we consider what those claims are. Specifically:

'The present author is satisfied that the assertions of NLP writers concerning representational systems have been objectively and fairly investigated and found to be lacking.'
(IV1, p 275)

    I realise that, strictly speaking, this claim is incontestable.  If we adopt a purely pedantic viewpoint, Dr Heap is on the face of it simply stating his opinion.  And if that is his opinion, then that is his opinion, and whether his opinion is 100% valid or utterly fallacious - or anywhere in between - is irrelevant.

    It appears to me, however, that the way in which he has presented his opinions indicate beyond a reasonable doubt that he was not offering, and never intended to offer, his views as a purely personal observation, but rather that he believed that he was/is presenting an objective, fair, and above all a factually accurate assessment of the material he was/is discussing.

    A second questionable aspect of the statement is the use of the words 'NLP writers', a label Dr Heap fails to define.  As I said earlier, it is certainly not acceptable to assume that everyone who writes about NLP-related topics is automatically qualified to present a fair and accurate version of those topics.  One book claiming to be about NLP, published only a few months ago, includes material on something the author calls 'the grey zone', which is not in any other book on NLP that I've read, and is certainly not an authentic NLP-related concept.  Likewise the author drags in 'Kundalini energy', 'Chakras' and 'spiritual awakening', though John Grinder has made it crystal clear that NLP was never intended to address any aspects of 'spirituality'.

    In short, if we were to accept this part of Dr Heap's statement as valid then he would be free to quote absolutely anyone who has written about NLP (as in Neuro-Linguistic Programming), regardless of how accurately their claims reflected the claims of the creators and co-developers, Richard Bandler and John Grinder.

    My objection, then, is not to Dr Heap holding any particular opinion, or his choice to accept any writer on NLP as authoritative regardless of whether their claims are true to the claims of the two genuine authorities on the subject.  What I question is his error in presenting his beliefs as facts, over a substantial period of time, when in practice these particular opinions are almost completely erroneous.
If Dr Heap's review is to have any value at all then it must be assumed that he meant it to be, and still regards it as being, tied directly to the claims made by Bandler and Grinder - and no one else.

    On this basis, given Dr Heap's overall conclusion as stated above, the following propositions must all be true:

  • The 'present author' (i.e. Dr Heap) must be genuinely qualified to make a valid assessment of:
    1. what the creators of NLP were asserting regarding representational systems,
    2. the accuracy of the understanding of those claims by the 'investigators' whose work was under review,
    3. the accuracy of the experimental findings, and
    4. the degree to which the experimental findings do or do not support the assertions that have actually been made regarding the representational systems by Bandler and Grinder.
  • The assertions made by Bandler and Grinder concerning representational systems must have been accurately understood by the investigators, and accurately reflected by their experimental designs.
  • The outcome of the experiments under review must have shown very little support for the relevant claims made by Bandler and Grinder.

The original review

    In his 1988, detailed review of various experiments on a couple of NLP-related techniques (Dr Heap has never discussed NLP itself in any of these papers - see point 1 in Evaluating the claims, above), Dr Heap included some 63 abstracts.  I have been able to obtain all but three of those abstracts - which are mainly of dissertations by candidates for a Masters or Ph.D. degree.  And the first point to notice, in light of Dr Heap's choice to use them as the basis for his evaluation for the whole field of NLP, is that they are of incredibly varying quality.  Some are clear and well presented.  Some are equally lengthy and detailed, but exhibit significant confusion as to the nature of 'NLP' and claims made about it.  Some are rather short on the kind of details needed, in my opinion, to make any kind of meaningful assessment of their worth.  And some are of such poor quality that I found myself wondering how their authors ever got on a Master's or Ph.D. degree course in the first place.

How objective is 'objective'?

    Unfortunately Dr Heap, in setting out his findings, is remarkably vague as to the import of the various experiments, saying only that 'The ... assertions of NLP writers concerning representational systems have been objectively and fairly investigated ...' (Heap 1988, p. 275).
But what does Dr Heap actually mean by 'objective' and 'fairly'?

    In practice it seems clear, to me, that Dr Heap signally failed to take a balanced view of the material he was reviewing, disregarding even the most obvious errors. For example:

  • Several of the projects tried to tie eye movements to preferred (or primary) representational systems (PRSs) - a task that is simply impossible since there is unlikely to ever be a one-to-one relationship between these two types of signal except in very unusual circumstances.  That is to say, eye movements (which occur in groups, or 'strategies') will inevitably outnumber rep' system signals (predicates, breathing patterns, etc.) several times over (see here for a fuller explanation).  If Dr Heap understood this fact, why did he not comment on it in his review?
  • Although Dr Heap acknowledges the fact that most of this research is done by relatively inexperienced researchers rather than by professionals, he seems not to attach any importance to this fact.  In practice, this element alone should have been enough to set alarm bells ringing and raise the question of whether the student researchers actually knew enough about NLP to design and execute meaningful experiments (see next point).  This seems to me to be particularly relevant given that Dr Heap has emphasised to me that he has never carried out any related research himself and is therefore, presumably, quite unaware of the difficulties of creating meaningful research in this particular area.
  • Most importantly, Dr Heap failed to spot the fact that Bandler and Grinder's recommended method of using the predicate matching technique made it practically impossible to use any single- or double-blind trial to test the efficacy of the process (see comments by Harris and Rosenthal in the evaluation of the NRC report).
  • Going by the abstracts, in the majority of cases the students made no allowance whatsoever for the influence of the observer/investigator effect, or, indeed, for their own limitations.  Thus although several studies apparently showed that none of the subjects were using a particular sensory system as their PRS (preferred representational system - see here for more details), instead of questioning the design and execution of their experiments.  In all cases bar one this was simply seized upon as 'evidence' that the NLP-related claims were unsupported.  Again Dr Heap seems to have been happy to accept all such self-serving bias without question or comment.
  • When it came to checking their results, some authors (e.g. Green (1979) and Rebstock (1980)) stated or at least implied that trained 'raters' or 'facilitators' assisted in their experiments.  But they do not say (in their abstracts), and Heap did not ask:
    • Whether all experimenters used 'trained observers'?
    • In what sense did each of these observers qualify as 'trained'?
    • Whether these observers had actually attended a training course in NLP techniques and concepts, or was simply given a book or two on the subject?
    • How extensive their training had been?
    • Who they were trained by?
    • And so on.
    Yet if he relied on the abstracts for this information then Dr Heap simply had no idea as to the answers to any of these questions. And neither do we.
  • Likewise several researchers made mention of previous experiments which allegedly failed to support NLP-related claims, but said nothing about whether this information had in any way affected their own expectations and/or behaviour.  And yet again Dr Heap pays no perceptible attention to a potentially crucial influence on the reliability of the experimenters' reports.

Think of a number

Next, in IV1, Dr Heap wrote:

'... in view of the absence of any objective evidence provided by the original proponents of the PRS hypothesis, and the failure of subsequent empirical investigations to adequately support it, it may well be appropriate now to conclude that there is not, and never has been, any substance to the conjecture that people represent their world internally in a preferred mode which may be inferred from their choice of predicates and from their eye movements'
(IV1, p 275.)

    Yet he has apparently never put a figure on what percentage of the abstracts he believed were pro, neutral or con in their support for NLP-related claims, nor which abstracts belonged in each group.  I would like to declare, by way of contrast, that I found that if all of the experimental abstracts in Heap's review were included (this excludes three reviews, three currently untraced abstracts and one non-NLP-related study - see below), some 50% of them appeared to support Bandler and Grinder's claims to a greater or lesser extent.

    To avoid turning this article into a mini series, the abstracts in Dr Heap's review which allegedly or actually offered at least partial support for the claims under investigation include:

Appel (1983); Beale (1981)*; Beck and Beck (1984); Bieber, Patton and Fuhriman (1977)*; Brockman (1980); Day (1985); Ellickson (1983); Ellis (1980); Falzett (1979); Frieden (1981); Frye (1980); Graunke (1984)*; Graunke and Roberts (1985)*; Hammer (1983)*; Hernandez (1981); Mattar (1980); Mercier and Johnson (1984); Owens (1977); Pantin (1982); Paxton (1980); Sandhu (1984); Schmedlen (1981); Shobin (1980); Wilimek (1979); Yapko (1981a) and Yapko (1981b).

    (Notes: I have omitted Kinsbourne (1972) from the list because although his findings seem to offer some support for Bandler and Grinder's claims, the work was carried out before NLP as such existed and has little or no direct bearing on the concept of representational systems.

    I am currently building a set of analyses of the individual experiments on my website which can be found at Cargo Cult Criticism.

The items marked with an asterisk are particularly interesting since their authors, and Dr Heap, claim that they contradict NLP-related claims when in fact they confirm them.
In the cases of Hammer (1983) and Graunke (1984), for instance, the experimenters correctly note that their subjects readily switched between rep' systems during an interaction.  Which they would have known, had they done adequate pre-project research, is exactly what Grinder and Bandler predicted:

'Our claim is that you are using all [representational] systems all the time.  You can shift from one to another.  There are contextual markers that allow you to shift from one strategy to another and use different sequences.  There's nothing forced about that.'
(Frogs into Princes, p.36. Italics as in the original)

    Again, I believe that in accepting the negative interpretations Dr Heap demonstrated his own lack of understanding of representational systems in general and PRSs in particular, and the eye accessing cues as described by Bandler and Grinder, and thus, I suggest, disproved his implied ability to make an accurate evaluation of the material he was reviewing.
So what went wrong?

    Having read IV1 and IV2, several times over, it appears to me that the objections raised all depend upon a misinterpretation of the claims made by Bandler and Grinder.  I suggest that Dr Heap has constructed a map of the territory which he treats as though it were absolutely correct.  And that he has done this despite having read Bandler and Grinder's books and having in front of him incontrovertible evidence that his map was inaccurate.  Thus, in IV2 he writes (in the final section before the references):

'These assertions [regarding PRSs, etc.] are stated in unequivocal terms by the originators of NLP ... it ought to be the case that writers refrain from, and editors of books and journals disallow, the presentation of such allegations as though they were well-established scientific facts ...'
(p 123)

    Question: Where do the originators of NLP actually take this approach?  Answer: Nowhere.  On the contrary, in Frogs into Princes in particular Bandler and Grinder positively and unambiguously reject any such behaviour.  As regards 'well-established scientific facts' Bandler and Grinder say:

'We have no idea about the "real" nature of things, and we're not particularly interested in what's "true",'
(p 7. Italics as in the original)

and a few pages later:

'As modelers we're not interested in whether what we offer you is true or not, whether it's accurate or whether it can be neurologically proven to be accurate, an actual representation of the world.'
(p 18)

    Do these really sound like the sort of statements someone would make if they are claiming to impart 'well-established scientific facts'?  Of course not.  Bandler and Grinder have always been clear that what they were studying fell within the realm of 'the study of the structure of subjective experience'.  A subject, it might be noted, that the Behaviourists chose to ignore entirely, and which even now is the focus of far more questions than answers in 'conventional' psychology.

    And here again, Dr Heap completely confuses the issue.  In his latest paper, for instance, he presents us with a 'mental map' which he calls The historical context of NLP.  It's quite interesting, but it is a map of what Dr Heap knows about that period - which has little or nothing to do with NLP.  Thus he makes an absolutely fundamental claim: 'NLP was one of a plethora of therapies that appeared from the 1970s onwards' with no supporting evidence whatsoever for this allegation.  On the contrary, he goes off at a complete tangent with no information whatsoever to tie NLP to the matters he is discussing.

    In practice, the 'historical context' of NLP has a great deal to do with topics such as Alfred Korzybski and General Semantics, along with John Grinder's professional interest in Transformational Grammar, and the wide open field of subjective perception and experience - and little or nothing to do with anything Dr Heap discusses in this part of his article.  In short, Dr Heap's 'history' is a perfect fit for his mental map, and totally ignores* all of the evidence that contradicts the story he wishes to tell.

    (*Caveat: Given the passages from Frogs into Princes quoted by Dr Heap I am assuming that he has read the whole book, including material from that book which I have cited in this article.)

Is NLP a form of therapy?  And if not, why not?

    At this point I think we are bound to consider what NLP really is, and isn't about.
Dr. Heap writes:

'... there is absolutely no question that the origins of NLP and its initial impact were in the field of counselling and psychotherapy'
(IV3, 2008. pp. 7-8. Bold font as in the original)

    This is near enough correct, as far as it goes, but after that we soon find ourselves misdirected by the implicit introduction of two related but illogical syllogisms that go something like this:

  1. Fritz Perls, Virginia Satir and Milton Erickson were all therapists
  2. Bandler and Grinder built NLP around their models of Perls, Satir and Erickson
  3. Therefore NLP is a form of therapy.


  1. Bandler and Grinder carried out certain therapeutic activities
  2. Bandler and Grinder created NLP
  3. Therefore NLP is a form of therapy.

    Problem: In neither case is Step 3 either logical or accurate.

    Bandler and Grinder were interested in modelling communication rather than therapy as such (see Frogs into Princes, pages 38 and 47, for example):

' in your work as a professional communicator ...'
'The proper domain, in our opinion, of professional communicators is process.'

    This brings in an important distinction frequently emphasised by Bandler and Grinder, between process and content.  The focus of all genuine NLP-related techniques is on process rather than content.  The focus of Heap's maps, and those of many other critics I've come across, is on content rather than process.

    To be specific, Bandler and Grinder were concerned with the process of excellent communication, not with the content of therapeutic communication in particular.  Although Bandler, and later Grinder as well, were involved with Gestalt Therapy (another reason why they weren't interested in creating some new form of therapy), the gist of the question Bandler posed to Grinder was: 'Explain how I'm doing what I'm doing in such a way that I can teach that to other people and so that I can communicate any information in a more effective way.

    By limiting his attention to the content of what was being modelled rather than the way in which the content was communicated, Dr Heap arrived at a completely erroneous understanding of what Bandler and Grinder wanted to do, and hence a completely erroneous view of what NLP in particular, and the field of NLP in general, were/are all about.  As a simple example:

  • Bandler and Grinder were studying communication;
  • Heap says that successful therapy depends on rapport existing between therapist and client;
  • How does a therapist create rapport except through their communications - both verbal and non-verbal?
  • But whilst therapy is always about communication, not all communication is about therapy.

NLP Modelling

    There are, as John Grinder points out, many valid forms of modelling, but only one procedure which qualifies as 'NLP Modelling'.  This was initially devised (in this context) by Richard Bandler and developed by Bandler and Grinder.

    It is crucial to any understanding of the whole field of NLP to know that NLP itself is this particular modelling technique and nothing else.  Everything else that people tend to think of as NLP is actually made up of NLP-related techniques and applications.  To answer Dr Heap's query as to how such a seemingly disparate collection ever came together under a single title, these techniques and applications were created or adopted/adapted either to support the NLP modelling process (as in the case of the meta model, for example), or because they were discovered as a consequence of using the NLP modelling process (as in the case of the fast phobia technique).
(Bandler and Grinder's books on the Patterns and Techniques of Milton H. Erickson, M.D. give some interesting and useful insights into the process of modelling Erickson and his work.)

    The important aspects of the NLP modelling technique are as follows:

  • The modeller collects as much information about the exemplar's behaviour, attitudes, beliefs, etc. as possible, with NO evaluation of the relevance/value of what is being gathered.  In modelling Fritz Perls, for example, Bandler even adopted a German accent and chain smoking in order to come as close as possible to how he had seen Perls acting in filmed therapy sessions and heard him on audio tapes.  He did this without pre-judging whether or not the accent and smoking were necessary to the success of Perls' communications.
  • Whilst collecting this information the modeller periodically tries to replicate the performance of the exemplar, using the latest version of their model.
  • When the modeller can replicate the results the exemplar achieves the modeller stops collecting information. As Bandler and Grinder put it in Frogs into Princes, (p 7):

    'We know our modeling has been successful when we can systematically get the same behavioural outcome as the person we have modeled.'

  • (By the way, being able to replicate the exemplar's results, i.e. getting the intended result, is an example of the NLP activity that seems to baffle many critics: 'Doing what works.')
  • The modeller then refines the model they now have, by testing each and every element, in order to remove whatever behaviour and so on isn't actually needed to achieve the required results.
  • And finally, the refined model must be recorded in such a form that it can be successfully taught to others.  Again the success of the process is determined by how well the trainees can reproduce, or exceed, the results produced by the original exemplar(s).

    Now, here comes the crucial information, as far as this discussion is concerned:

    As a student at the UC Santa Cruz Bandler started out studying mathematics and computing, and ended up studying psychology.  During this time he subsidised his studies working in the warehouse of a local book company.  This led to him getting the job of editing a book on Fritz Perls, which included transcribing a number of tapes of Perls' therapeutic sessions.  According to John Grinder, Bandler was so skilled at absorbing aspects of Perls' work with Gestalt Therapy that he ended up being able to use Gestalt Therapy even more effectively than Perls himself.  Indeed, it is said that when he ran out of material for the book, Bandler was able to finish the job by writing from within his internalised model of Perls!

    Bandler also spent some time with family therapist Virginia Satir, managing the sound and recording systems when she was running training sessions and demonstrations. Here, too, he built a highly effective model of her techniques, even though he allegedly spent most of his time reading books and was therefore only peripherally aware of what Satir was doing.

    Thus Bandler, and later Bandler and Grinder, had detailed access to the work of two outstanding therapists, not because therapy was what they particularly wanted to study (note Bandler's initial disinterest in what Satir was doing), but because Bandler's work happened to create those opportunities.  And as Grinder pointed out to me, getting close to people who are recognised by their peers as genuine models of excellence within their chosen field is not an opportunity that presents itself very often.

    The third exemplar was psychiatrist and clinical hypnotist, Milton Erickson.  And again the connection was established almost completely by chance.  Bandler and Grinder had by this time advanced to the point where they were able to co-author a book on family therapy with Virginia Satir, thus demonstrating the effectiveness of their modelling of her work.  The only limitation at this stage was that Perls had died not long before Bandler edited the book referred to above, so Bandler and Grinder were on the lookout for a another 'model of excellence' whom they could compare with their models of Perls and Satir.

    John Grinder lived close to, and knew, the British anthropologist Gregory Bateson (who was also working at the UCSC at the time), and it was Bateson who recommended that they (Bandler and Grinder) go to Arizona to study Erickson and his work.

    In short, Bandler started getting acquainted with the work of a couple of outstanding therapists - examples of 'excellence' in their chosen activity - entirely by chance.  Once the process had started it was logical that Bandler and then Bandler and Grinder would continue to model therapists until that part of their project was complete.  Likewise their own work in the role of therapists was a case of testing the accuracy of their models and not because they were therapists, or wished to create some new form of therapy.  And they passed on their knowledge to people who would be involved or at least interested, in psychotherapy because they were in a position to use the models in their own work and thus test the final stage of the NLP modelling process.
Nor is there any ambiguity or confusion on this point.  As Bandler and Grinder stated very clearly in the seminar which was the basis for their 1979 book Frogs into Princes:

'We [Bandler and Grinder] call ourselves modelers.  What we essentially do is to pay very little attention to what people say they do and a great deal of attention to what they do.  And then we build ourselves a model of what they do.  We are not psychologists, and we're also not theologians or theoreticians.'
(p 7. Italics as in the original.)

    Let us proceed, then, in the sure and certain knowledge that the claims that NLP is a kind of therapy, are completely untrue.  NLP is simply a specific modelling procedure.

Back to '88

    So what about those experiments? What are the representational systems (usually abbreviated to 'rep' systems') really about?  And why were the experimenters of the 1970s and '80s so far wide of the mark?

    There are 5 rep' systems - visual, auditory, kinaesthetic, olfactory, and gustatory.  They approximate to the five basic sensory systems, though in NLP-related jargon the term 'kinesthetic' is applied to feelings in general - tactile feelings, visceral feelings and emotional feelings.  (If this seems untoward in any way, please remember that NLP-related techniques and concepts are about what is useful rather than about conventional 'truths'.)

Primary Representational Systems

    In all honesty I must confess that I initially made the same mistake that Dr Heap, and many others, have made about the concept of primary representational systems.  That is to say, I thought (based on secondary sources) that Bandler and Grinder were saying that each of us has a preference for just one representational system which we tend to stick to.  In fact the matter is somewhat more nuanced.  What Bandler and Grinder actually say is:

'How many here now see clearly that they are visually oriented people?  How many people see that? How many people here feel that they are really kinesthetically oriented people in their process? Who tell themselves that they are auditory? Actually all of you are doing all of the things we're talking about, all the time. The only question is, which portion of the complex internal process do you bring into awareness? All channels are processing information all the time, but only part of that will be in consciousness.'
(Frogs into Princes, p 34. Italics as in the original text)

    and a couple of pages later:

'Our claim is that you are using all systems all of the time.  In a particular context you will be aware of one system more than another.  I assume that when you play athletics or make love you have a lot of kinesthetic sensitivity. ...'
(ibid, p 36. Italics as in the original text)

    Notice, here, that Bandler and Grinder are claiming that the process whereby one rep' system predominates is context specific and not, as so many experimenters have apparently assumed, a fixed-for-life phenomenon.  In fact several researchers amongst those included in the review were testing how long what they assumed was a primary representational system remained primary.  Given John Grinder's statement that people can switch rep' systems as frequently as every 30 seconds (approx.), these students had clearly missed the point entirely.  The fact that Dr Heap accepted these abstracts as relating to valid experiments argues that he shared this misconception.

    Moreover, on the day before I sent this article off to Dr Heap, and long after I initially wrote this section, I discovered that Einspruch and Forman (1985) made this same observation in their criticism of Sharpley's (1984) review.  This means that, had Dr Heap read the full text of their paper at the time, he would have known exactly what he needed to know in order to have avoided his basic error regarding the nature of representational systems.

Incidentally, I suggest that the fact that Einspruch, Forman and I are all 'singing from the same hymn-sheet' - over 20 years apart and without prior reference to each other's writing - is further evidence that Bandler and Grinder have maintained the same ideas on this subject over the intervening period.

    Next, Bandler and Grinder also noted that some people exhibited not one but two or more representational systems as the primary representational system, with the qualification that:

'... we tend to use one or more of these representational systems as a map more often than the others.'
(The Structure of Magic 2, p 8)


'... a person may have more than one most highly valued representational system, alternating them. This is common in people who are incongruent in their communication ...'
(op. cit., p 26).

Given that most of the subjects in most of the various experiments under review were university students, this last point is especially relevant.  What we now know, though we didn't in 1976, is that due to certain features of brain development during our teens and early 20s, this is a time when incongruency is, for many young people, more or less a fact of life.

The core mistake

    This, then, is the core mistake made by most of the experimenters, and in the reviewing process - the expectation that people will use a single representational system (the one the experimenter/reviewer takes to be the person's primary representational system) in preference to all others come what may.

    This is not correct, and Bandler and Grinder, as we've already seen, were not making that claim. Not even in the late 1970s.  Like most NLP-related techniques, the favouring of one or two representational systems over the others has always been viewed as something that is context-based.  As an example, some years ago there were several series of a UK TV programme called Masterchef, presented by Loyd Grossman, in which trios of amateur chefs prepared meals in the studio which were judged by Grossman, a celebrity and a professional chef - it being a knock-out contest in format.  (The current version of the series is "but a pale imitation".)

    The interesting thing in relation to the contextual nature of rep' systems was that all through the preparation cycle the professional chefs talked about the food almost exclusively in terms of smell.  But once the dishes were presented for judging they switched to evaluating the food in terms of visual appearance and taste.  (Note, this applied to almost every chef over several series.)

    This clearly illustrates, as some of the reviewed experimenters noticed, how easily and naturally people can and do switch from rep' system to rep' system according to the requirements of the moment.
Which is why Alan Hammer (1983), was actually confirming Bandler and Grinder's findings when he reported that people's verbal and non-verbal signals need to be tracked ('calibrated' in NLP jargon) and responded to throughout an interaction - not just at the start.

    It is my contention that Dr Heap, and most of the student researchers whose work he reviewed, had a responsibility to find out what Bandler and Grinder were really claiming.  Instead they opted to rely on their own misinterpretations.  Had they double-checked their maps with a genuinely authoritative source, had they read Bandler and Grinder's books with more care, then, I believe, virtually all of the studies would have supported the claims for the relevant NLP-related techniques.

    (For what it's worth, times seem to be a'changing.  Just a few weeks ago I received an enquiry from the Chair of a department in a Mid-West College who is about to embark on research of the eye accessing cues model, asking for any observations I might have.
By the way, I'm not suggesting that researchers should write to me in particular.  In fact I have passed the Professor on to someone who has a particular interest in this topic and who has been working with Bandler for over 20 years.)

    It is a simple fact that the field of NLP is now in its fourth decade and is going from strength to strength.  It extended way beyond the original group not because of 'brilliant marketing', as one online/book author has claimed, but because it was given credence by people such as Elizabeth Loftus*, Daniel Goleman (in Psychology Today), Gregory Bateson et al.

    *(Professor Loftus (1982) referred to Bandler and Grinder's account in Frogs into Princes of implanting entire false histories in people as a way of making them feel better.  For example, working with people who had been fat all their lives, they successfully implanted false childhoods in which they had grown up thin.)

    Dr Heap argues that NLP [sic] should be amenable to rigorous scientific testing.  But the NLP-related techniques (which is what Dr Heap and other critics are actually talking about) were never offered as methods or models which would always work, for every NLPer, with every client, in every context.  In practice, NLP in general has succeeded because the related techniques work enough of the time, for enough people, in enough contexts, to make them generally useful.  Since this is a known feature of the NLP-related techniques it is clearly nonsense to expect to get useful results from subjecting them to the 'scientific method'which is predicated on 100% consistent behaviour in the materials being tested.

    There is too little space, here, to develop an in-depth response to the charge that NLP should be investigated in a 'scientific' manner.  Instead I would refer anyone interested in the topic to read Dr. Liam Hudson's book The Cult of the Fact (1972).  Dr Hudson was a professor of psychology and his book details the reasons why it inappropriate to apply scientific-style investigative methods to psychology.
As to NLP in particular, Bandler and Grinder have always said, if what you're doing isn't working, do something else instead - and keep varying your behaviour until you find something that does get you the results that you're after.


    In his IV1 paper, Dr Heap concluded his report with these words:

'This verdict on NLP is, as the title indicates, an interim one. Einspruch and Forman (1985) were probably correct in insisting that the effectiveness of NLP therapy [sic] undertaken in authentic clinical contexts of [sic] trained practitioners has not yet been properly investigated.
'If it turns out to be the case that these therapeutic procedures [sic] are indeed as rapid and powerful as claimed, no one will rejoice more than the present author.  If however these claims fare no better than the ones already investigated then the final verdict on NLP will be a harsh one indeed.'
(p 276)

    Unfortunately, as I believe I have demonstrated here, it appears that Dr Heap:

  • Was prepared to simultaneously hold the contradictory views that (a) the experiments he reviewed were 'fair and objective' but failed to support the claims he thought had been made for NLP AND (b) that what he (mistakenly) called NLP therapy had NOT been properly/adequately investigated.
  • Took little or no account of the many obvious opportunities for error, and actual errors, in the various experimental abstracts.
  • Had no idea that NLP itself was just a particular form of modelling.
  • Was also under the impression that there was little or nothing more to 'NLP' than primary representational systems, sensory predicates and the eye accessing cues.
  • And carried out such a restricted version of the research he is generally thought to have performed (online citations make claims such as 'Dr Heap carried out a systematic review'), that he specifically cited, but apparently failed to read, the one paper that would have explained why his review was so profoundly in error.

    (To be fair, the bibliography in IV1 makes it clear that Heap only read the abstracts of the various papers.  Unfortunately many people who have subsequently cited Heap's review seem unable to understand how very limited these abstracts are as regards the amount and nature of the details they supply.)

Thus in IV1 and IV2 I believe I am right in saying that he doesn't make even a passing reference to:

  • The Meta Model
  • Meta Programs
  • Goal setting
  • Anchoring
  • Calibration
  • Various rapport-building techniques such as mirroring and matching, and cross matching
  • Presuppositions (NLP presuppositions in particular and and linguistic presuppositions in general)
  • Chunking
  • The use of metaphors
  • NLP-related problem resolution
  • and so on, and so on.

    Even in IV3, Dr Heap's only recognition of the true breadth of NLP seems to be his brief comment that he is concerned by the way Noam Chomsky's ideas have been incorporated.  Is he aware, I wonder, of the fact that Grinder had already co-authored a college level textbook on Transformational Grammar before he started working with Bandler?  And is he aware that the NLP-related 'meta model' is an adaptation of Chomsky's model, not an attempt to co-opt a carbon copy of Chomsky's ideas as an NLP-related technique?

    Here, again, and I promise you that I do not say this lightly, I would argue that all three of Dr Heap's papers demonstrate how little understanding he had of the subject, and how seriously underqualified he has been, both then and now, to offer any kind of competent opinion on the material he reviewed.


Bandler, R. and Grinder, J. (1975) The Structure of Magic I. Palo Alto, CA: Science and Behaviour Books, Inc.
Bandler, R. (1992) Magic in Action. Capitola, CA: Meta Publications.
Bradbury, A. (2006) Develop your NLP Skills (3rd edition). London: Kogan Page.
Dilts, R. (1994) Effective Presentation Skills. Capitola, CA: Meta Publications.
Dorn, F.J. (1983) The effects of counselor-client predicate use in counselor attractiveness. American Mental Health Counselor's Association Journal, 5, 22-30. (Designated 1983a in Dr Heap's references.)
Einspruch, E.L. & Forman, B.D. (1985) Observations concerning research literature on neurolinguistic programming. Journal of Counseling Psychology, 32, 589-596.
Frith, C. D. (2007) Making up the Mind. Oxford: Blackwell Publishing.
Gazzaniga, M. (1998) The Mind's Past. Berkeley: University of California Press.
Grinder, J. & Bandler, R. (1976) The Structure of Magic II. Palo Alto, CA: Science and Behaviour Books, Inc.
Hammer, A.L. (1983) Matching perceptual predicates: Effects on perceived empathy in a counseling analogue. Journal of Counseling Psychology. 30, 172-179.
Heap, M. (1988) Neurolinguistic programming: An interim verdict. In M. Heap (Ed.) Hypnosis: Current Clinical, Experimental and Forensic Practices (pp 268-280). London: Croom Helm.
Heap, M. (1989) Neurolinguistic programming: What is the evidence? In D. Waxman, D. Pedersen, I. Wilkie & P. Mellett (Eds.) Hypnosis: The Fourth European Congress at Oxford (118-124). London: Whurr Publishers. (see )
Heap, M. (2008) The validity of some early claims of neuro-linguistic programming, Sceptical Intelligencer, 12, 6-13
Hudson, L. (1972) The Cult of the Fact. Jonathan Cape, London.
Jones, D. & Motluk, A. (2008) How to get exactly what you want. New Scientist, May 10.
Juhnke, Gerald A., Coll, Kenneth M.; Sunich, Michael F. & Kent, Ronda R. (2008) Using a modified neurolinguistic programming swish pattern with couple parasuicide and suicide survivors.  The Family Journal, Vol. 16: pp. 391-396. October
Laborde, G. (1983) Influencing with Integrity. Mountain View, CA: Syntony Publishing.
Laborde, G. (1988) Fine Tune your Brain. Mountain View, CA: Syntony Publishing.
Loftus, E.E. (1981) Memory and its distortions. In A.G. Kraut (ed.), The G. Stanley Hall Lecture Series (pp. 123-154). Washington DC: American Psychological Association
Sharpley, C.F. (1984) Predicate matching in NLP: A review of research on the preferred representational system. Journal of Counseling Psychology, 31, 238-248.
Sharpley, C.F. (1987) Research findings on neurolinguistic programming: Nonsupportive data or untestable theory? Journal of Counseling Psychology, 34, 103-107.