by Drs. Jaap Hollander
In the late 19th century a man named Thomas Edison propagated a new system whereby light could be produced using electricity. Around the same time a certain Bell wanted to set up a wild thing he called a telephone line. Fortunately, the general public was warned by academic experts, who at an early stage were onto the glaring scientific invalidity of these novelties. Edison and Bell were exposed as intellectual swindlers whose ideas could not pass the test of scientific scrutiny. Erasmus Wilson, professor at the University of Oxford, wrote at the time: ‘When the World Exhibition in Paris is over, light by electricity will also be finished and will never be heard of by anyone again.’ A British parliamentary committee which had been formed to examine Edison’s light bulb, agreed wholeheartedly with the professor. In 1878 they drew up the following conclusion: ‘Edison’s ideas do not deserve the attention of people who think in scientific and practical terms.’ Just like the lightbulb, the telephone was also quickly unmasked to show its ramshackle scientific underpinnings. In 1863 a certain Joshua Coopersmith even was arrested as a confidence trickster on fraud charges, when he tried to raise money to set up a telephone system. The science journalist of the ‘Boston Post’ wrote: ‘Well informed individuals know that it is impossible to have the human voice transmitted through a wire, and that such an invention, if it were feasible, would be of no practical value at all.’
In Holland, where I live, scientists have, over the last ten years, repeatedly ran down NLP in the media. It is likely that you have the same experience in your country. A Dutch professor of psychology called Drenth for example, once slated NLP on national television during a discussion with my partner Anneke Meijer. When Anneke explained the principles of modeling, the professor stated that ‘modeling had nothing to do with NLP.’ He also claimed to have made a ‘thorough scientific analysis’ of NLP. Apparently the principles of modeling which are described in just about every book on NLP, had escaped this thorough analysis. Drenth further claimed that: ‘NLP is a fake name, with which people are using in an attempt to suggest a scientific basis to the marketplace.’ And here I guess we have the prototype of the scientific condemnation of NLP: a respected scientific authority voices a definitive rejection based on a quick review of a few NLP-books written by motivational speakers who claim NLP can do everything people have ever wanted. Of course the motivational speakers who write these kinds of books have no scientific proof of their statements. I guess that most of them don’t even have much of an idea what the scientific method even is, precisely, nor do they claim they do. I doubt that you’ll be very surprised to hear this. But the interesting fact is, that precisely the same goes for the professor. In the above case of professor Drenth, there is not one single scientific study that supports his firmly postulated rejection either. Nor did he try to cite even one single piece of research. And unlike the motivational speaker, the scientist does allege to knowing the scientific method. This truly puzzles me. How can a scientist make firm statements that aren’t backed up by any scientific data? A salient detail is that at the time professor Drenth was the chairman of the Academy for the Sciences, a highly respected Dutch institution which had indicated that it wanted to set up a system for reporting scientific fraud. One of the offences that could be reported was ‘slandering colleagues without grounds’. The strange thing is, that if we had wanted to expend the energy, we could have reported the professor to his own Academy….
In my personal practice NLP offers scores of useful concepts, models and tools that other, more ‘scientific’ approaches do not offer. But when something works, especially in the area of psycho social processes, more often than not it is too complicated to research. And when it is being researched in the scientific, hypothesis testing, statistically analyzing manner, the phenomenon is usually reduced to the point where in daily practice the results are useless. Unless, and this is what often happens with scientific data, these results are generalized in an unscientific way. I may find, for instance, that with certain groups of male college students a certain muscle relaxation technique works better than an open conversation about their problems. But when I work with an overstressed female senior manager, how can I use this scientific facts? In order to be able to say that muscle relaxation is going to work better for her, I would have to generalize from male to female, from old to young, from student to manager and similarly for hundreds of other important differences. Of course I cannot do that and still be scientific. In science generalizations are supposed to be tested. In other words: pragmatic and empirical criteria tend to clash. It is impossible in daily practice to work, even if I should want to, in a manner that is truly scientifically supported. There will always be unwarranted generalizations because there will always be differences between my client and the subjects in the experiments. And yet I work and yet I gain experience and make generalizations about my experiences. So when I write these experiences down, can some scientist may call me a swindler?
Science is a Belief System Too
A Dutch advertising monthly called ‘Intermediair’, which is sent to the academic community for free, published an article on NLP which had a large headline saying: ‘In fact we are dealing with a belief system’, as if this was a definitive disqualification of NLP. For me however, the world is held together with belief systems. So I agree completely that NLP is a belief system. But what is science? When a scientist claims disparagingly that NLP is only a belief system, is he then seriously implying that the scientific method is based on anything else than beliefs? Isn’t science essentially based on the belief that (a) there is a truth, that(b) it is good to uncover this truth and that (c) statistically evaluated experiments supply the most useful test of this truth? At least that was the situation when I was a university student. If it is different today, then a great revolution in the philosophy of science has completely escaped my attention. Scientists believe that statistically evaluated experiments produce useful information about the truth. As far as I know, they have never proven this. But most scientists show an obstinate kind of subcultural ethnocentrism. They seem to be saying: ‘Our belief is the truth and your belief is only a belief’. This is reminiscent of the cultural anthropologists from the late nineteenth and early twentieth century, who contrasted the ‘superstitions’ of native peoples with their own ‘Christian faith’: “You religion is superstition, while mine is true faith”. Some scientists would retort that science, even if it is a belief system, is at least a culturally accepted belief system, one about which we have consensus. But isn’t one of the traditional duties of science to challenge culturally accepted beliefs?
Grinder wants to study the exceptions NLP is not a scientific theory. NLP models and techniques are not based on a systematic set of experimentally and statistically supported hypotheses, nor are they tested in this manner. John Grinder has stated, from the beginning of NLP, that he fundamentally disagrees with the scientific method, in the sense that he does not want to study the averages; he wants to study the exceptions. The history of NLP starts with modeling exceptionally talented or experienced people in order to transfer their unique abilities to others. NLP came into being through generalizing from the mental and emotional possibilities of exceptional individuals. This experiences was filtered through a loosely gathered set of concepts borrowed from transformational grammar, the semantic theories of Alfred Korzybski, cybernetics, social learning theory, hypnotherapy, gestalt therapy, etc. Words like ‘deep structure’, ‘representational system’, ‘ecology’, ‘modeling’ and ‘feedback loop’ were taken from the various frameworks mentioned above. And this is where some confusion originates. Bandler and Grinder borrowed their concepts from scientific systems. Yet the way in which they used these concepts is not scientific, e.i. didn’t follow the rules of hypothesis testing called for by the scientific method. It is understandable that this confuses many scientists. Some of them are now saying that the linguistic and neurological terms in NLP only serve as metaphors to represent psychological processes. The majority of NLP authors, including Bandler and Grinder, have never claimed otherwise.
Another source of confusion in this area is the absence of a centralized certification agency. In the present state of the field, any crackpot can use the term NLP and – to aggravate matters further – can even show an official looking document proving that he is a certified expert. As soon as someone like that starts writing books, it becomes exceedingly difficult for scientists to determine what ‘NLP’ claims or doesn’t claim. In Holland there is at least one book that presents NLP as a ‘new scientific method’ and one other book that describes both Bandler and Grinder as ‘professors’ (Grinder was at one time an associate professor, but Bandler never was). Scientists can hardly be expected to distinguish between one brand of NLP and the other. So, given the fact that concepts borrowed from science are used within a non scientific framework plus the fact that a minority of NLP writers call NLP ‘scientific’, we can expect scientists to protest. I suggest that NLP authors explain that they find the use of certain scientific concepts handy as a filter to organize their experience, but that they use these concept in a loose manner which does not imply that they use the scientific method. Furthermore I think any NLP author should refrain from calling NLP scientific.
You Can Get Very Wet in a Scientifically Dry Area
Some scientists claim that the models and methods of NLP are not scientifically supported and therefore they are worthless and the public should be warned. This connection between thrush and value however is a philosophical choice rather than a scientific fact. Pragmatist philosophy as propagated by William James for instance, makes another choice (it’s good if it works). As I mentioned before, no scientist has ever proven that scientifically supported generalizations are more useful than other generalizations. The choice for scientific truth therefore, is a philosophical choice. I think scientists often confuse their philosophy with their facts. The beliefs that (a) there is a truth, that(b) it is good to uncover this truth and that (c) statistically evaluated experiments supply the most useful test of this truth, are just that: beliefs. They may be very useful or even noble beliefs in some contexts, they may be long surviving and highly accepted beliefs, but nonetheless they are beliefs and as such they are not inherently better than other beliefs. From a position of confusing their map with reality, scientists often imply that it is better to use scientifically proven methods. There are not only philosophical, but also some practical problems involved in this, the usefulness of scientific facts. The most important of which being (a) that group statistics are not directly translatable to predictions about individuals and (b) that scientific experiments reduce complex interactions to an extent where conclusions are become irrelevant to actual practice.
Let’s first take a look at the issue of statistics and their use in practice. Knowing that problem X can be solved with method Y in 46.3 percent of the cases, is of little importance when the person sitting in your office belongs to the remaining 53,7 percent. And if there was an other method which works for only 0.001 % of all people with problem X, then that is the right method if someone belongs to precisely that one-thousandth of a percent. Even though this method could be claimed to be almost totally ineffective scientifically speaking. If I walk outside and it is raining, I will get wet. Even if it can be proven without a shadow of a doubt that in 83.3% of The Netherlands the weather is gloriously dry. I would be getting wet in a country that can be claimed quite scientifically to be dry. Therefore, the announcement that a ‘scientifically sound’ method is being used, has a suggestological meaning at most. The fact that 50% of a certain group is successful with method X, unfortunately does not mean that a certain individual also has a 50% chance at success. Group statistics are not translatable to predictions about individuals. So if any method should claim to be a ‘scientifically proven’ method or that ‘scientific research has shown the method to be effective’, this is almost automatically a misleading and unscientific statement.
A Student With a Piece of Paper
I have already mentioned two problems in the relationship between science and practice. On was the difficulty with generalizing statistical group data to individuals and the other was the uncertainty about science being a good guide for practice, since scientists have never proven that scientifically supported methods work better than other methods. Now I turn to a third problem with the usefulness of scientific truth: the trade off between depth and precision. In scientific research one tends to gather very precise information about a very limited range of phenomena. In the case of people’s thoughts, feelings and actions, a scientist can only gather precise, statistically processable information about a small part of a process. The process is thereby strongly reduced, often to the point where the relationship with the original issue to be researched becomes quite dubious. In quantitative research the conditions to which the researched entities are subjected, must be exactly the same for each entity. All test subjects in a certain group must be given exactly the same procedures under exactly the same circumstances, otherwise one doesn’t know what one is measuring.
This is already a very difficult proposition when it comes to human experience. First of all, to paraphrase John Grinder: when I hit a golf ball with a club this would be pretty much the same for different balls. When I hit a person with a club however, he may process this event in many different ways. The same stroke will be different things for different people. But an even more serious problem with the scientific method is that socio-psychological processes are often reduced to such an extent that the relationship to actual practice becomes highly debatable. For instance, when a scientific publication on psychotherapy mentions that method X has a Y percent success rate, what does that mean? If a scientists says: “We administered method X to the subjects”, we might ask the metamodel question: “How precisely did you administer it?”. This question is not nearly as shallow as it may seem, because the term ‘method X’ is already a huge nominalization. Oftentimes the actual process behind this nominalization is nothing more than a student reading out instructions for method X from a piece of paper. Maybe this student is doing this for the first time in his life, or else after a very short training. As a comparison condition he then reads from another piece of paper the instructions for method Y. Can you imagine how this would work if a student were comparing, say, reimprinting with core transformation in this way? Who would be surprised when subsequently little difference is found between methods X and Y? “Scientific data proves that there is no difference between method X and method Y”. The specification “when they are read from a piece of paper by third year psychology students” never appears in the abstract of the research paper, which may on the other hand contain impressive tables of numbers and complicated statistical tests. “The hypothesis that ‘method X’ is more effective than ‘method Y’ was not supported by this study”. Maybe it just means that students reading instructions are not very effective as change agents. What strikes us again and again, however is that scientists who subsequently talk about ‘method X’, will never talk about the complex equivalence.
NLP as a Series of Introspective Field Experiments
NLP focuses on the individual. Apart from critics, the academic world also has advocates of NLP. At various Dutch universities NLP-courses have been given, be it sporadically and to a limited degree (Rotterdam, Utrecht, Leiden, Twente). Educators, pedagogues, management administrators and psychotherapists have been introduced to NLP on a wide scale and they often interweave NLP-concepts in other frameworks. Dozens of students have graduated on research into NLP or have written extended essays about aspects of NLP. In NLP World several articles have been published describing experimental, statistically evaluated research into NLP. There are also at least two doctoral dissertations here in Holland which centered on the evaluation of NLP. The academics who are involved with NLP are not only fascinated by the practical applications of NLP, sometimes they also regard NLP as a step forward in the development of psychology. A step which elaborates on introspective research traditions from the previous century, which were represented by Wilhelm Wundt and William James, amongst others. Each time NLP is used, it can be considered an introspective field experiment: intentional changes are evoked and evaluated outside the laboratory. Because NLP techniques are standardized reasonably well, we can treat this kind of data as experimental results. Because of the large numbers of individuals doing these experiments – the thousands of practicing NLP professionals—a type of statistically reliable information comes into being. Be it that the reporting on this information is extremely loose and informal. here, by the way, lays another interesting opportunity for the NLP community. With today’s surveying techniques on the internet, it should be relatively easy to collect sink-or-swim data: who uses which techniques how often for what?
Those who who reject data of this kind of research on formal grounds, are forgetting that long before the term ‘Appellation Control’ was invented, Bordeaux wine of excellent quality already existed. We see science as an interaction between two major elements: generalization and specification; inductive and deductive reasoning. From observations of and musings about complex events in life, people make generalizations about the causal relationships within these events. Thus, ideas about essential elements, their properties and their relationships or ‘laws’ are being formulated. This is a creative process that is a necessary part of science. The next step in science is to test these relationships in experiments and to evaluate the results statistically. NLP could for scientists be a rich source of generalizations that have already stood – to some extent and in a loose sense – the test of time. Most NLP concepts, models and techniques exist in a sink-or-swim environment. The ones that work keep being used and the ones that don’t work well enough go to that place up high where they can rest and drink beer with the other techniques whose time has come and gone. This sink-or-swim test might turn out to be a wonderful pretetst for scientists, where promising candidates are defined for further research. For the practitioner who wants working technologies, the sink-or-swim test might be just as useful if not better than scientific research.
Five Recommendations to Improve the Relationship Between NLP and Science
1. Scientists should refrain from judging phenomena they have not researched, in order to stay within the limits of their scientific competence.
2. It would be wise for NLP authors to emphasize that they use concepts from scientific models as filters and metaphors rather than elements of traditional scientific hypothesis testing.
3. Remember that science is a belief system just like other belief systems, be it a more generally accepted one than NLP. Scientists have never proven that the scientific method will lead, immediately or eventually, to better results than other methods.
4. Be aware that statistical data about groups of people cannot be translated directly to individuals, and that the phenomena researched are often reduced to the point of irrelevance. Claims that methods are ‘scientifically proven’ are misleading.
5. NLP can be seen as an extensive series of introspective field experiments, generating a crude kind of statistical data: does a model or a technique sink or swim? Scientists can use this as a pretest for practically relevant hypothesis. NLP practitioners might formalize this process some more.