32 - Did Bandler and Grinder Really Understand NLP?


"The 'Science' of NLP" - Did Bandler and Grinder have it right all along?


Due to the ways that search engine algorithms interpret search requests, you may or may not have ended up at the right page.  To help you to get to the information you are really after, here are some related pages in this site which you might want to visit as well as, or instead of, this one:

  • FAQ #2 - Describes what "NLP" is.
  • FAQ #3 - Describes where "NLP" and the NLP-related techniques came from.
  • FAQ #20 - Describes some of the limitations of the NLP techniques.
  • FAQ #21 - An example of critism of "NLP".  Two school textbook authors use NLP-type linguistic techniques to criticise "NLP".
  • FAQ #22 - A discussion of research of NLP and NLP-related subjects.  Includes references to over 100 such studies which have produced positive findings.
  • FAQ #27 - A detailed rebuttal of the wildly inaccurate article on "NLP" in the so-called Skeptics Dictionary.
  • FAQ #28 - Following up on FAQ #27, this "multipart" FAQ shows how poor research has meant that a dependence on flawed and even unequivocally false information about the FoNLP (field of NLP) has been commonplace amongst academic critics for over 20 years.  The subsections include details of the infamous 'reviews' by Sharpley (1984, 1987) and Heap (1988, etc.).
  • FAQ #32 - Describes exactly why research into preferred representational systems and predicate matching, which makes up an overwhelming majority of the negative "evidence", is based on an absolutely fundamental flaw.


(*   The FoNLP is the "field of NLP".  This includes NLP itself, a specific modelling technique + the NLP-related concepts, techniques and applications + training in NLP and/or any of the NLP-related concepts, techniques and applications.)

If there is one aspect of "NLP" that stands head and shoulders above every other criticism it is the plaintive cry that "NLP" hasn't been scientifically validated.

(The sometimes implicit, sometimes explicit corollary - that the authentic FoNLP* has been scientifically disproved - is illogical (a "false dichotomy") and untrue.  Anyone who believes that this corollary holds enough water for a gnat to bathe in is cordially invited to visit FAQ #28 (and its sub FAQs), which deals with the claim in detail.)

Over the last decade or three we have been increasingly bombarded with messages from certain members of the scientific community, telling us that "science" is the only valid measure of what is meaningful or "true".

There is nothing new about this idea.  It is a "retread" that has its roots at least as far back as the 1920s, when a group of philosophers in Vienna, known as Logical Positivists, announced that "scientific knowledge is the only kind of factual knowledge and all traditional metaphysical doctrines are to be rejected as meaningless." (Encyclopedia Britannica online - see References).

On this basis, when Bill whispers "I love you" to Betty, Betty should point out that this is a meaningless statement since "love" cannot be empirically verified (i.e. weighed, measured or captured in a test tube).

Or should she?

According to Professor A.J. Ayer, who championed the adoption of Logical Positivism in the UK in the 1930s, "Logical positivism died a long time ago".  And he said that back in the 1980s.  Yet here we are, a decade into the 21st century, and the proponants of a remarkably similar belief system - known as materialist reductionism (and sometimes as reductionist materialism) - are trying to sell us this discredited pig in a poke all over again.
Once again we are being told to believe that only science has the answers - or will have, one day.
Maybe not today.  Maybe not tomorrow.  Maybe not next year ....

But why would we want to re-adopt a failed belief at all?
Should we in fact be asking:

"Has the FoNLP been scientifically validated?"

Or would it be much more useful to ask:

"Is it sensible to believe that all of the elements of the FoNLP must be amenable to scientific testing?"

It seems that the second question is more to the point than the first, because the critics themselves can't agree on an answer.  Even Dr. Christopher Sharpley, whose two reviews (1984 and 1987) have provided a deeply flawed basis for so many later critics, actually decided that:

... perhaps NLP principles are not amenable to research evaluation.  This does not necessarily reduce NLP [sic] to worthlessness for counseling practice.  Rather it puts NLP [sic] in the same category as psychoanalysis, that is, with principles not easily demonstrated in laboratory settings but, nevertheless, strongly supported by clinicians in the field.
(Sharpley, 1987.  Page 105)

On this occasion it seems Sharpley was at least looking in the right direction, though for the wrong reason.

conspicuous example of an element of the FoNLP which is not open to evaluation by means of standard empirical research methods.  It is also the very technique and concept which had supposedly been investigated in most of the experiments covered in Heap and Sharpley's reviews.

This is why we need to take a careful look at a background paper prepared for the Enhancing Human Performance committee in 1987.  A paper, by the way, that seems to be entirely based on Sharpley's 1984 review, as far as "NLP" is concerned.  That is to say, not a single NLP-related text is mentioned in the list of references for this paper, let alone any material by the co-creators, Richard Bandler and John Grinder.
We can, however, find some crucial errors in Sharpley's first article repeated in this paper, especially Sharpley's claim that there were:

... three NLP methods of determining the PRS [page 242] ... eye movements, verbalizations, and self-report [page 239]
(Sharpley, 1984)

And yet - for all its faults, this paper throws some much-needed light on the question asked in this FAQ: "Is it sensible to believe that all of the elements of the FoNLP should be amenable to scientific testing?"

I Can See You!

Discussing one of the experimental reports reviewed by Sharpley (1984), these authors comment that:

... there are features of this experimental design that leave open the possibility that expectancy effects may be responsible for some of the positive results.  The most striking characteristic is that both the interviewers and the experimenter were not blind either to the hypotheses under study or to the experimental condition of the subjects.  Because they had assisted in the assessment of subject's PRS, the interviewers knew whether or not they were matching their subjects' PRS.
(Harris and Rosenthal, 1987.  Referring to Falzett (1981))

If these comments are correct, and they are, this means that it is impossible to carry out a truly "blind" or "double blind" study of the authentic NLP-related predicate matching technique.

The authors of the background paper - on Interpersonal Expectancy Effects and Human Performance Research - were Monica Harris (then a Ph.D. candidate at Harvard, now a Professor of Psychology at the University of Kentucky), and Robert "Pygmalion in the Classroom" Rosenthal (then a Professor of Psychology at Harvard, now Distinguished Professor of Psychology at the University of California, Riverside).  And having spotted the "fly in the ointment", at the end of this section of their paper they offered their idea for how to get round the problem:

"Blind" testing
The use of single- and double-blind test procedures are intended to block participant expectations which might otherwise affect the test results.
Single-blind - the person(s) administering the test or the subjects (not both) do not know whether a given subject is in the test group or the control group.
Double-blind - neither the person(s) administering the test nor the subjects know whether a given subject is in the test group or the control group.

Based on the above considerations we recommend that in planning the research on the NLP construct, more studies should be conducted that use standardized audio- and videotapes, or that assess subjects' PRS prior to the interaction and keep reviewers blind to the subjects' PRS.  Studies where the ongoing use of predicates is tracked by the interviewer are too vulnerable to other changes in interviewer behavior; too much is left uncontrolled, making it harder to conclude that PRS matching vs mismatching is the causal agent.

Given these considerations, it is very interesting to note that a recent review of the NLP research shows that the strongest support for PRS matching was obtained in tracking studies (Sharpley, 1984. p. 246).  More total PRS matching occurs in the tracking studies, which might account for the stronger results obtained.  However, studies that used standardized videotapes or that assessed subjects' PRS prior to the interview do not support the NLP model [sic] (Sharpley, 1984).

This lack of support leaves open the very real possibility that interpersonal expectancy effects are responsible in part for the positive results found in the predicate tracking studies.
(Harris and Rosenthal, 1987.  Page 26)

Notes:   The text has been split in three to make it easier to read onscreen.  The actual wording is exactly the same as in the original document.
(A collection of all of the background papers is available online (read only) at no cost.  See References section below.)

In light of the two authors' seriously distorted view of "NLP", their analysis of the situation, and their recommendations, actually make very good sense.

Unfortunately, however, their view of "NLP" is indeed completely distorted, and as a consequence, from a "real life" point of view, the recommendations are completely inappropriate and positively self-defeating.

Don't Panic!   Don't Panic!

The key phrase in the above quote is: "too much is left uncontrolled" (see the section of FAQ #22 which addresses the difficulties raised when trying to deal with what are called confounding and extraneous variables).

These authors recommend using the techniques which provide the best opportunity of keeping track of the "causal agent" - the "true cause" of whatever "effects" are observed.  And their concern about the possible effects of "interviewer behaviour" is entirely justified.
Taken in isolation the entire recommendation is justified.
But we cannot take the recommendation "in isolation", only in context.  And the fact is that Harris and Rosenthal appear to have made their recommendation in the fallacious belief - presumably borrowed from Sharpley (1984) - that both eye movements and self-reports are genuine options, sanctioned by authentic "NLP" guidelines.

In practise, however, they are recommending the two fantasy options whilst ignoring the one genuine claim made by Bandler and Grinder.  Because the co-creators of the FoNLP specified only one means of establishing someone's PRS, both before the amendment to the description of a PRS:

In order to identify which of the representational systems is the client's most highly valued [preferred] one, the therapist needs only pay attention to the predicates which the client uses to describe his experience.
(Bandler and Grinder, 1976.  Page 9)

And after:

The "representational system" is what's in consciousness, indicated by predicates.
(Bandler and Grinder, 1978/1979.  Page 28.  Italics as in the original text)

(See Whatever Happened to Preferred Representational Systems for a short explanation of the authentic, albeir modified, NLP-related view of PRSs as it was from at least as far back as early 1978.)

Where's the Reciprocity?

The recommendation that: "studies should be conducted that use standardized audio- and videotapes" (Harris and Rosenthal, 1987.  Page 22), actually makes no sense when considered in the context of the authentic NLP-related technique.  The interviewer's part of each interaction can only be usefully taped if we know in advance which modality is "preferred" by the subject being interviewed.  But this possibility was ruled out by the early part of 1978, leaving a situation where the interviewer must match/mismatch predicates in response to the interviewee's on-going use of predicates.  If the interviewer's part of the interview is all pre-prepared and stored on tape then the essentially reciprocal nature of the authentic predicate matching technique would automatically be excluded.

Thus when Harris and Rosenthal say that:

... studies that used standardized videotapes or that assessed subjects' PRS prior to the interview do not support the NLP model (Sharpley, 1984)

they are overlooking the rather vital point that this "NLP model" is Sharpley's own creation which they have uncritically adopted, not an authentic NLP-related model.

The irony is that it doesn't matter that Harris and Rosenthal have wandered off down a blind alley, because their conclusion is entirely acceptable:

This lack of support leaves open the very real possibility that interpersonal expectancy effects are responsible in part for the positive results found in the predicate tracking studies.
(Harris and Rosenthal, 1987.  Page 26)

We Call Ourselves Modelers

In order to understand this apparent paradox we need to turn to Bandler and Grinder's own explanation of what they were doing.  That is to say, they were never interested in making a "scientific" case for the efficacy of the NLP-related techniques.  Which is why, at least as far back as 1978, they were telling their seminar delegates:

We call ourselves modelers. ... We are not psychologists, and we are also not theologians or theoreticians.  We have no idea about the "real" nature of things, and we're not particularly interested in what's "true."  The function of modeling is to arrive at descriptions which are useful.  So if we happen to mention something that you know from a scientific study, or from statistics, is inaccurate, realize that a different level of experience is being offered you here.  We're not offering you something that's true, just things that are useful.
(Bandler and Grinder 1978/1979.  Page 7)

And again:

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.
(Bandler and Grinder 1978/1979.  Page 18)

The point, then, is this:

The authentic NLP-related techniques have been modelled from people who are widely recognized as being excellent exponents of whatever behaviour was being modelled, or based on such modelling and tested in real life situations.

If a given technique works enough of the time, on enough clients, in enough situations then it is likely that it will also be useful for a significant number of other people.  And now comes the tipping point, so to speak:

The co-creators of NLP weren't, and still aren't, looking for scientific or academic validation, because they aren't basing their offerings on having a knowledge of "why" something works - or why something "ought" to work.  In this case, for example, if using the predicate matching technique works - is followed by the required results, to a reasonably consistent extent - then why it triggered those results, be it directly or indirectly, is really not an issue.

The point may be illustrated by a story from the 1978 seminar, the edited transcript of which became Frogs into Princes:

When I entered the field of communication, I went to a large conference where there were six hundred and fifty people in an auditorium.  And a man who was very famous got up and made the following statement: "What all of you need to understand about doing therapy and about communication is that the first essential step is to make contact with the human you are communicating with as a person."  Well, that struck me as being kind of obvious.  And everybody in the audience went "Yeahhhh! Make contact.  We all know about that one."  Now, he went on to talk for another six hours and never mentioned how.  He never mentioned one specific thing that anybody in that room could do that would help them in any way to either have the experience of understanding that person better, or at least give the other person the illusion that they were understood.
(Bandler and Grinder 1978/1979.  Page 7)

Conclusion - Did Bandler and Grinder Really Understand NLP?

The FoNLP is built on modelling because its co-creators compiled observations about what people who were excellent in some respect did when they were being "excellent".
It is, in my perception, an error to suppose that the FoNLP is about why the techniques succeed, whether they have theoretical underpinnings which make them scientifically valid, and so on.  Indeed, the Harris and Rosenthal paper shows very clearly why the technique that has come in for most attention over the last three decades couldn't sensibly be investigated in a "scientific" manner.

This seems to indicate that Bandler and Grinder had a very good understanding of what they were doing.  Based on my studies to date (May, 2010) of a number of well-publicised studies of what the various authors thought of as "NLP", it seems to be the critics of the FoNLP who have been lamentably wide of the mark.  And apparently all for one reason - they haven't done the research needed to produce useful results.


Bandler, R. and Grinder, J. (1990), Frogs into Princes  Eden Grove Edition, London.
Originally published in the US by Real People Press, Moab, Utah, 1979.
This is an edited transcript of a seminar that took place in Denver, Colorado in March, 1978.

Commission on Behavioral and Social Sciences and Education (1988) Enhancing Human Performance: Issues, Theories, and Techniques, Background Papers (Complete Set).  Accessed at http://www.nap.edu/openbook.php?record_id=778&page=R1 in May 2010.
(Note: the papers are independently numbered, starting from page 1 each time.)

Encyclopedia Britannica, online version.  Accessed here on May 15, 2010: http://www.britannica.com/EBchecked/topic/346336/Logical-Positivism.

Falzett, W. C. (1981), Matched versus unmatched primary representational systems and their relationship to perceived trustworthiness in a counseling analogue.  In Journal of Counseling Psychology, 28.  Pages 305-308.

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.