The disastrous state of NLP training

Steve Andreas—with even more valuable input from Connirae than usual



I’ve spent the last 35 years of my professional life — and much of my personal life — learning, developing, and training high quality NLP. Recently I saw a video in which someone was teaching the swish pattern in a way that greatly weakened it. Looking around a bit, I found 16 videos of the swish online. I was dismayed to find that none of them taught it as originally presented, and all but one made the same very fundamental mistake, as well as many others. The fundamental mistake is equivalent to replacing the engine in a Lamborghini with a hamster wheel. Other mistakes are like putting wagon wheels on it.

These mistakes show not only a widespread lack of ability to learn and follow the steps of the pattern, but also a lack of understanding of the principles underlying each step. In this article, I’ll review the 16 videos (which provide you with sensory-based experience) and point out the mistakes. I hope this can add to the understanding not only of what to do, but why to do it, which is essential to the field’s integrity and progress. But first, a little history.



The swish pattern is a rapid way to change any troublesome habit or other unwanted response, so it has a very wide range of applications. The swish was developed by Richard Bandler in the early 1980’s, and was first published in Using Your Brain for a Change (chapter 9) in 1985, over 30 years ago. Connirae and I taught it and used it extensively over the next couple of years. During this time we accumulated a lot of experience of when it didn’t work, or only partially worked, and we had to figure out what we needed to do to correct that. In Change Your Mind—and keep the Change (chapter 3) we included many additional details, including how to design a swish in the auditory modality. A case example, with follow-up, appeared in our later book, Heart of the Mind (chapter 17). We also produced a video on the swish in the early 1980’s, including two demonstrations, one of the standard size/brightness swish with nail-biting, and also an auditory “designer” swish with a woman who went “ballistic” when her daughter used a particular tone of voice. These sources provide a rich description of all the different essential elements in the process, the principles underlying each element, and specific examples of how to make the pattern work.


The “standard” swish using size and brightness, (or size and distance)

This swish is often used to teach the basics of the pattern. Often it’s demonstrated with nail-biting, because it’s easy to identify what the client will always experience immediately before the problem behavior or response — their hand has to move up toward their face just before biting their nails.

After identifying this cue image, the next step is to elicit a desired self-image of “the evolved you of the future, for whom nail-biting is simply no longer an issue.” The client is asked to see themselves in a dissociated self-image, much like a 3-D portrait. Seeing this positive self-mage provides strong motivation, engaging unconscious processes to develop ways to become like the self-image. This creates a direction for change at the more powerful identity level, in contrast to only selecting a specific replacement behavior.

This desired self-image works best when it is seen without context or background, and not doing any specific behavior. Any background would tend to limit the scope of generalization to that context, and picturing a specific behavior would limit the change to that behavior.

In the next step, the client is asked to see the cue image (of hand coming up to face) big and bright, and somewhere in that image to see a small dark dot containing the desired self-image. Then the client is told to allow the desired self-image to very quickly become big and bright as the cue image shrinks and becomes dark. This links the cue image to the desired self-image, so that any time they are in a situation that used to trigger the unwanted behavior, they will immediately see the self-image.

After a break state, the client is asked to exchange the images again, repeatedly, with a break state in between, to make sure the direction is always from the cue to the self-image. After 7-10 repetitions, the cue image often becomes insubstantial or disappears, while the self-image becomes prominent, so this is one way to test the intervention. Asking the client to bring a hand up to the mouth is another good way to test, and real-world follow-up is best of all.

The dissociated self-image provides powerful motivation to change, without specifying how the change will occur, which is left to the client’s unconscious processes. The change is usually instantaneous, and the client usually isn’t consciously aware of any specific behavioral change. Often they’re “just a different person” who wouldn’t even think of biting their nails. If the self-image were associated, that would assume that the client had already become it, so there would be no motivation to change, only a self-delusion that change had already happened.

This standard swish usually works well as a very simplified introduction to the pattern. But it makes a lot of assumptions, and omits many important details. For instance, it uses size and brightness because for most people those two variables will increase the feeling response to any image, whether pleasant or unpleasant. Ideally we would test this, and often use other visual (or auditory) submodalities that elicit the strongest response for a particular client. The standard swish doesn’t include training in how to elicit a dependable cue image when there is no obvious external image like the hand coming up to the face. There is no testing to be sure that the client has a strong positive response to their self-image, or training in what to do to increase their response. There is no mention of searching for a positive intention of the problem state, or what to do about satisfying it. There are no suggestions for how to “trouble-shoot” when the test at the end indicates that the process didn’t work. There is nothing about using other visual submodalities, or how to design a swish in the auditory or kinesthetic systems, etc. Despite all these limitations, the standard swish is a good way to teach the overall structure of the process; training in the additional details can be added later.


The importance of the desired self-image

The most significant mistake (which appears in all but one of the video presentations of this pattern that I reviewed) is that instead of a desired self-image, the client is asked to see themselves doing a specific substitute behavior. Using a specific behavior instead of an evolved self-image is an error that also appears in Robert Dilts Encyclopedia of NLP. “Form a mental image of herself engaged in the behavior she would like to do instead of smoking.” Since Dilts has written in great detail about the power of identity in change work, this fundamental mistake is particularly surprising. A consciously chosen image of a specific behavior provides only one option — which may not fit very well — in comparison to the infinite variety of possible choices that a positive self-image can generate unconsciously.

If you create a sequence from a problem behavior to a more desirable behavior, that is essentially chaining with submodalities, not a swish pattern. That can work, but it won’t be nearly as dependable or powerful, for several reasons.

A positive self-image is far more powerfully motivating and generative than an image of a different behavior. For example, an image of yourself flexing your fingers (instead of biting your nails) just isn’t as intensely motivating as seeing “the you that you would be without this problem.”

As the cue image for the problem behavior becomes smaller and darker, the unpleasant feeling decreases at the same time that the self-image becomes larger and brighter, and the pleasant feeling increases. This smooth analog transition can be diagrammed using a rectangle with a diagonal dividing it into two triangles, one green and one white, to show how this creates motivation away from unpleasantness (green) and toward pleasantness (white). (In the diagram, time goes from left to right.)


In contrast, the diagram for a simple chain from one state to another, would show one (black) state ending, and another (yellow) state beginning, an abrupt digital change that occurs in a moment in time. That can work, but the connection isn’t as solid. The diagram above is like a lap joint; the diagram below is like a butt joint. Anyone who has worked in wood, metal, or fabric will tell you that a lap joint is always far stronger and more lasting.


In Dilts’ article on the swish, he credits the power of the method to the submodality changes (size and brightness), but this is only partly true. These submodality changes can only amplify the existing feelings in response to the two images. If those feelings are not very intense to start with, there is little for the submodality shifts to amplify. It is really important that the feeling in response to the cue image is unpleasant, and that the feeling in response to the self-image is as strong as possible, so it’s important to test. “See that image of the evolved you for whom biting nails is simply no longer a problem; how desirable is that image?” and notice the nonverbal response as well as the verbal. If the response is weak, you need to do whatever you can do to intensify it before proceeding. Often this will involve resolving incongruence, or amplifying other submodalities.

In the 16 videos of the swish that I’ve watched recently there is quite a lot of variation, and only one of them (sort of) used a desired self-image. Every video either modified the pattern in ways that weakened it, or left out important steps. Only a few of them included how to test to find out if it had worked or not. If a test didn’t work, the only remedy offered was to repeat the process a few more times. None of the videos include any follow-up that would indicate how successful the process actually was. The specific details are in the video reviews below.


Reviews of videos

I found videos on YouTube by searching using the terms “swish” “swish pattern,” and “swish pattern NLP.” I may have missed a few, but the huge variation in them clearly shows how much they differ from the swish pattern as originally developed and taught by Bandler. I encourage you to watch the videos before reading my comments, because my comments will make a lot more sense if you already have the sensory-based experience provided by the video. That also gives you an opportunity to notice errors, and compare what you observed with my comments. If you have only a little time to watch videos, I suggest watching the first four, since they provide a wide range of examples. Quite often the language used in these demonstrations is ambiguous and imprecise, but there are so many examples of this I won’t usually comment on them. (6:10)

1. Michael Carroll describes the swish as being useful for “a minor behavior” which implies erroneously that it can’t be used for more intense difficulties. He role-plays the pattern in order to demonstrate it, walking himself through the steps. He uses nail-biting as the problem, and size and distance as the submodalities. His description of how to select the cue image could be more precise, but it’s basically correct — an image of what he dependably sees out of his own eyes just before the problem behavior/response, namely fingers moving toward the mouth.

Then he says, “Think about what it is that you’d like to do instead,” and goes on to specify “staying within the frame” of the cue image. This asks for a behavior (in the same context) rather than a desired self-image (without a context).

He suggests that he might pick “flex your fingers or do something with your hands,” as the new behavior to replace nail-biting. This is a specific behavior chosen by the conscious mind, in contrast to the evolved self-image. Michael is a primary sponsor of John Grinder and his “New Code” method, in which a decision about a new behavior is completely turned over to a client’s unconscious mind. So it’s particularly puzzling that this principle of relying on unconscious selection is ignored in this demonstration, despite its being an explicit part of the original swish pattern.

In the middle of self-demonstrating this change process, Michael switches from biting nails to cracking knuckles as the problem to be resolved. Changing the content “midstream” isn’t ideal for teaching, but isn’t a problem with the method itself.

Then he says and demonstrates “see themselves flexing their hands,” as the new behavior, and then to shrink that image “very small, and send it far out to the horizon.” Then he has the close cue image go quickly out to the horizon, as the outcome image quickly comes in. His verbal instructions only mention distance. The corresponding change in size is presupposed, but not mentioned explicitly. For the swish to be effective, it’s best to explicitly utilize two submodalities together. (8:50)

2. In what must be a quite old video, Tony Robbins teaches a group, also using nail-biting as the behavior to be changed. He is very expressive, and clear in his instructions to get the cue image. His instruction for the outcome image is a mixture of specific behavior and self-image: “What you want is a big bright colorful picture of myself looking good, lecturing, and being proud of myself, and also noticing that my nails are great, too.”

Then Tony demonstrates the transition. He uses size, and gestures with his right hand grabbing the self-image and bringing it up so that it gets larger and explodes and breaks through the cue image, without mentioning a second submodality. Grabbing the small image is a nice kinesthetic addition that will make it more powerful for some clients. Tony says, “I get the experience of feeling good about myself,” in response to the self-image. Tony summarizes the overall pattern nicely as, “I don’t need to do this (the problem behavior), this (the outcome image) is who I really am,” continuing to emphasize the positive self-image.

The basic process here is correct. However, he uses only one submodality, size, and the self image “breaks through the cue image, rather than simply getting larger as the cue image gets smaller. In his desired self-image, he mixes in a specific behavior, i.e. “a picture of myself … lecturing.” This context tends to limit the unconscious from generalizing the change to other contexts where it would be useful.

“Being proud of myself,” is certainly a motivator for Tony, but it’s not ideal, because pride (one of the 7 deadly sins) is one half of a very troublesome polarity (the other half is shame) that compares the self to others. Pleasure or satisfaction is a much better motivator, because it’s an evaluation that doesn’t require a comparison with others — and when many people use the word “pride” loosely, that is what they really mean. The detailed structure of pride and shame is beyond the scope of this article; if that interests you, see chapter 10 of my book, Transforming Your Self.

Next is a short staged video vignette of a father and daughter in which Tony gives a summary of the steps to the swish. In this description, Tony makes a significant error in each step. Tony says the first step is to “see how you’re behaving now,” a dissociated image. The cue image will only work if it’s what you’ll actually see at the time (your hand coming up to your face).

In the second step Tony says to “think of how you want to behave — get a clear picture of what you’d like to behave like in the future.” Here he asks for a behavior, rather than an image of the person you want to become — the person who would no longer have this problem.

Tony suggests making the background of the cue image red (“stop”) and the background of the outcome image green, (“go”) which is a cute addition, but one that has a problem. The green, signifying “go” for the desired self-image might make it more powerful for some people. However, the cue image works best when it is a close match for what the client will see in the problem context. Very seldom will the problem context actually have a bright red background, so that will make it a poor match for the real world, and less likely to work as a cue.

Next Tony walks the group through using the swish to change a fear, asking them all to “get a picture of the fear” for the cue image. That language is ambiguous at best; “see what scares you,” or “see what you see just before you get scared” would be more precise.

Then when he asks for the outcome, it’s a self-image as in his first description — “seeing yourself as you would be, free of that fear, but also seeing yourself the way you’d be if you were totally proud of yourself; if you felt really good about who you are as a person, something that would really motivate yourself.” This instruction is basically fine, except for including the emphasis on pride.

Tony then walks the group expressively through the size transition, break state, repetition and testing, and then talks about how the self-image has a generative ripple effect, “It didn’t just make me feel different about my fingernails. . . . it now makes me think to be like the person who’s elegant, who I respect, who I want to be. So I found myself doing a lot of other things I wasn’t doing before, like just picking up things around the house. Before I’d just let it go, but the person I pictured in my mind wouldn’t just let it go, so I’d handle it.” This is a very nice description of the generative impact of the desired self-image, which goes far beyond any specific alternate behavior chosen by the client’s conscious mind. (7:38)

3. Akistis guides the viewer through the process, starting out with, “In your upper left (gesturing with her right hand) I want you to imagine the thing you want to improve,” which is more than a bit vague, since “thing” could indicate a partner, a garden, or an automobile. Then she clarifies, saying, “See a holographic image of yourself going through some negative emotions.” This instruction is a little more specific but has two problems. One is that she asks the viewer to be dissociated, rather than associated. The second problem is that she asks the viewer to see multiple emotions, rather than select one.

Next she gestures with her left hand (the viewer’s upper right) and says, “Imagine yourself in something that you know for certain, something that you do well, and that you are confident in.” Then she says to bring that first image over to the same space as the desired image, put it in front of it, and “integrate those feelings” (gesturing with both hands in a swirling motion) “until the stronger sense takes over the weaker sense.” This is much more like a clumsy version of the “visual squash” than a swish.

Then she asks the viewer to imagine someone on your left “who you dislike, who makes you feel weak, disempowered, in a situation where they’re getting you upset.” (3:42) “And now I want you to imagine on your right hand, right here, see the positive feeling, your feeling when you’re self-confident, the faces of people who really make you smile — think about those people’s smiles, let those feelings, let those juices flow, let those positive hormones come out, and just feel all that positive energy that you have in exchange with those people.” This is verbatim; I couldn’t make up such a mishmash of instructions. What does it mean to “see a positive feeling” or to “let those positive hormones come out”? And how do you combine seeing yourself confident with seeing other people’s smiles? Then she “integrates” the two pictures “So the strong image is washing over and disintegrating the old image.” (integration becomes disintegration!) Akistis then follows with a lot of new-age feel-good bogusities, such as, “the image will no longer find anything to grab upon in your own mind,” and “going to the next level.” This has got to be the worst example of teaching the swish I have ever seen, though the next one is a close second. (This is a download link; 34:14)

4. Mark Hayley demonstrates what he calls a “kinesthetic swish” with a workshop participant. In this “director’s cut” the demonstration is interrupted so that Mark (and an interviewer) can discuss the process. If you want to bypass the discussion, the demonstration alone can be viewed in the following short segments: 5:04-7:00, 10:55-12:33 16:30-17:39 Mark first asks the client to:

  1. Identify a bad feeling that she has experienced repeatedly.
  2. Notice the location of this feeling in the heart area of her chest.
  3. Mark pulls it out of her chest and asks her what it looks like.
  4. She sees it as a “yucky green rotating blob.”
  5. Mark tells her to shrink it down and make it black and white.
  6. Mark elicits a sudden feeling that dissipates quickly, which is located in her throat area.
  7. Mark tells her to bring the shrunk-down black-and-white image back into her body in the location where she has the sudden feeling (in #6 above).
  8. Which then dissipates quickly and automatically out of her body.

Whatever you think of this intervention, it bears no resemblance to the swish. The only submodality change that links all these steps is location; there is no second submodality as in the original swish pattern, and no self-image (or behavior) representation.

Curiously, step 3, pulling the feeling out of her body, is much the same as the last step in which the feeling dissipates out of her body. Why not just pull the feeling out of her body at step 3 and fling it across the room and be done with it, as many masseurs or “body workers” do in order to get rid of accumulated tension or “bad energy”? If you want to learn how to do a true kinesthetic swish, read this article published 18 years ago: (3:17)

5. NLP and Hypnosis Training New York (animation). The unwanted behavior is eating donuts; the cue is a hand reaching for a donut. The new behavior is working out (image of woman in gym clothes with barbell and exercise ball). “Fling the picture out into the horizon. It will get far away, small, and dark.” (three submodalities rather than two, which is OK.) “When it comes back, it comes back as your ideal self.”

In the swish, the cue image gets smaller, etc., at the same time as the self-mage gets larger, etc., so the motivation dynamic, “away from negative, toward positive” is active throughout the analog transition. In this example, the two are only connected at the horizon, like the butt joint I mentioned earlier, making both the motivation and linkage much weaker. In addition the positive image is of a behavior (working out), instead of being an identity image (the you who you would be without this problem). (3:31)

6. (animation) Shawn Carson applies the swish to dealing with a difficult person “Burt.” “How would you like to feel? If you could be anyone when dealing with Burt, who would you choose to be?” (image of superman in cape) This mixes in a bit of the “new behavior generator” pattern to get an image; there is no instruction to make this into a self-image, rather than an image of someone else. The superman image also makes it a “magical solution” rather than something real, suggesting that the only way to deal with Burt would be to have super-human powers. When the self-image is unreal, it won’t be very motivating. If the client believes in something unreal, that is a delusion. Either way, it’s not useful.

The next instruction is: “Take this ideal you, the new you, and shrink the picture down until it’s small enough to fit on a postage stamp. In your imagination, stick the stamp with the picture of the new you on Burt’s forehead. Have the picture of the new you spring out from Burt’s forehead. The new you appears life-sized in the space in between you and Burt.” In these instructions the image of Burt does not change, so it will continue to elicit the unwanted response. “Repeat several times.” There is no mention of a break state in between repetitions, which can result in a yo-yo effect, instead of the single direction from the problem image to the desired image. “Step forward into that new you, as you step forward to greet Burt. Feel how good it feels to be this new you.” This implies that the change has already happened, so there is no motivation to change further. The self-image needs to stay dissociated in order to elicit motivation to change. (10:50)

7. Anthony Beardsell works with a live client, Michelle, who has difficulty responding in a meeting with superiors when she is asked a question. He asks “Is there a state or mood that you’d prefer to be in at that point in time?” This asks for a resourceful feeling, rather than either a behavior or a self-image. Michelle says, “I’d like to feel as confident as I do when I talk to them on a one-to-one basis.” Anthony uses her feeling as a lead to elicit images, both of which are in the same location close in front of her and associated. Although it’s not entirely clear, she appears to be seeing what she saw when she had the feelings. Neither is an image of her evolved self. He has her step out of the desired image to use in the swish, using size, brightness, and distance (three submodalities rather than two). In order to use distance, he has to ask her to move the positive image away from her, which is inelegant at best; if he had only used size and brightness, that would not have been necessary. Michelle soon reports that she has only an empty frame for the cue image, so the chaining appears to have been successful, but it isn’t likely to be generative. (13:45)

8. David Shepard begins with a clear description of the dynamics of the swish, before eliciting the client’s problem state, which is fear of sliding when he’s on a motorbike going around a turn. After identifying the cue picture of beginning a turn, David asks, “How do you want to be?” asking for an alternate behavior, not a self-image. The client replies, “I can feel the tires completely locked into the road,” indicating an associated experience. David asks, “As you think about that, do you have a picture?” and then “For a moment, step into that picture, so that you’re looking out of your own eyes.” This assumes that the client previously had a dissociated image, which I think is probably not the case. David then asks the client to change the size, brightness, color, speed of the movie image, to intensify the response — all of which would have been more useful if it were a self-image. Then he asks the client to step out of the picture again, and uses this as the picture to swish to.

After a couple of swishes using size and brightness, the client reports (7:33) “It’s very fast, it goes about half-way when you tell me to open my eyes; it takes a bit of time to get up there,” with expressive gestures indicating great effort. This slowness and effort could be just inexperience with the transition, but it could also be an indication of incongruence that needs to be taken into account.

There is a real danger in the “solution” David invites his client to construct. If you’re going around a turn fast on a motorbike, and you have a dissociated image of yourself going around the turn, you’ll have reduced access to the feeling experience of being on the motorbike, which is an important part of what you need in order to turn safely. Although an evolved self-image is also dissociated, the result of using a desired self-image will be useful, since this change will occur almost instantaneously — long before he gets on a motorbike again.

After several more swishes, when David asks the client to get the first picture back (8:10) he replies that he can “sort of” see it, that it is indistinct, and (8:40) that the second image doesn’t expand to fill the whole screen. “It doesn’t go ‘whist,’ it goes ‘uurngh’ ” and again he gestures expressively indicating great effort. Since this slowness and effort persists, it is more likely to indicate incongruence, and I think it likely that it is because of the danger of having a dissociated image while going around a turn. David urges him to do it faster, and tells him that the first picture will eventually disappear. Since he has made this suggestion, when the picture later disappears, it’s not a good test of the process, since it might only be response to the suggestion. After some more swishes, David tests by asking him to imagine being on a motorbike on a turn and the client feels very different, and as if the bike is “on rails,” so it will keep to the turn. While it’s possible that being “on rails” is only a metaphor, I worry about that. In fact, a motorbike is not on rails; it will definitely slide if it goes too fast on a turn, so this could establish a dangerous delusion. (32:28)

9. Jevon Dangeli spends the first 8 minutes or so offering a number of frames, both general and specific to the swish, before asking for a volunteer and eliciting the cue image, which is a “picture of things not going the way I want them to go, me being ignored.” When Jevon asks her to identify what in the external world triggers this internal picture, she says, “an aggressive look” in their eyes. Then he says, “Tell me about that peaceful confident state (which she mentioned earlier) that you’d like to have,” and she says, “feeling relaxed, no tension in my body.” “When you’re feeling relaxed, there’s no tension, and you’re peaceful and confident, what image represents that for you?” When she asks, “An image of myself, or?—” Jevon says, (14:27) “Both — anything you want,” and she responds, “(I see) myself walking on a beach.” Not only is this an image of a specific behavior, rather than an evolved self-image, but if the swish works so that she sees herself walking on a beach in response to an aggressive look in someone else’s eyes that’s not likely to be useful in the real-world context.

Then he sets up the cue image, with the beach image shrunken to a speck in the middle of it, and whooshes that image out beyond the horizon, and come back as the beach image. This uses only one submodality, distance, and the two images are only connected at the horizon. This makes both the motivation and the linkage much weaker than if one image increased at the same time as the other decreased. He repeats the whoosh a number of times, each time following with hypnotic suggestions about the cue image changing. Jevon tests using the cue image, and in several future scenarios, and the client reports feeling peaceful and confident. He ends the demonstration at 24:07 and then responds to questions from the group. (7:17)

10. Mel Cutler works with a very responsive client who has trouble organizing papers on his desk. Mel appears to be reading most of what he says from notes. “How would you like to feel or act instead?” asks for a feeling or behavior rather than an evolved self-image. “Step into that picture so you’re looking through your own eyes, . . . adjust the submodalities of that new positive picture to make it more desirable than ever before, etc.” Stepping into this image isn’t necessary, and isn’t in the original swish, but since he later has the client step back out of the picture to use for the swish, it’s not likely to be a problem. He tests at the end, getting a 9 out of 10 and does the swish a few more times, until the client reports only seeing the desired image. (8:15)

11. Pip Thomas (edge NLP) asks client for an image of a “current negative behavior that you have in a certain context,” and then asks her to “step out of the image so you can see yourself in it.” Then she asks the client to get an image of the “new behavior you want to have in that set of circumstances,” rather than a self-image. She asks the client to adjust the color, brightness, volume, feeling, to make her response more compelling, and then to step into the picture. Both images are of behaviors; the old behavior is dissociated, the new behavior is associated, the reverse of the standard swish. Then she does a size/brightness swish, “Old behavior up on the screen, new behavior in the bottom right-hand corner, ready, one, two, three, swish—big and bright, old behavior gone.” Pip tests by asking client to imagine being in the problem context, and she feels “completely different.” (8:51)

12. George (, audio with graphics) explains that the pattern is used to replace an unhelpful feeling with a “good emotion that you’d like to feel instead.” (No behaviors or self-image) “Next we’ll attach those emotions to some colors and shapes. As you think of that negative emotion, think of the situation where this comes up, and focus on this red circle (on screen, left). . . . Now think of that good feeling that’s going to replace the bad feeling, focus on this blue circle (on screen, right).”

Then George repeatedly tests the connection between the colored circles and associated feelings, using circles that start small, and grow to be the same size as the previous circles. Then the intervention appears on the screen: a small red circle appears on the left and grows to its previous size, followed by the blue circle that grows and covers the red circle. The growing red circle will increase the old feeling (rather than decrease it) and after that the blue circle will increase the replacement feeling. The two changes are only connected at the point where the blue circle first appears, a butt joint. This sequence is gradually sped up until it is very fast. (6:00)

13. Alex (therapytipstv) asks the client, who procrastinates doing homework, “And how would you like to be in that situation. I’d like you to build an image of you doing your homework in exactly the way you’d like to do it,” which is an alternate behavior, not a self-image. Size is the only submodality mentioned when he has her do the “swish.” After some repetitions, he has her check with the homework image, and rehearses her twice in the future. Each time she reports feeling calm and relaxed. (9:22)

14. The Indian accents make this one very hard to hear — and the background music makes this worse. The practitioner asks for a preferred behavior, and then asks about positive intent, and checks to see if the new behavior satisfies the positive intent. This is fine, except that eliciting the positive intent ought to come before the target image, so that it can guide its selection and elicitation. He adds in additional pieces of having the client compliment himself, and see himself “achieving all his goals.” I gave up trying to hear about half-way through the video. (11:41)

15. Terry Elston works with a client who has some kind of trouble (hard to hear her) on Sunday evening. He asks her, “What would you like to do instead?” asking for a behavior rather than a self-image. She describes a positive feeling, using the word “possibilities,” with nice expressive expansive gestures. Then Terry talks her through a transition from “six o’clock to the positive feeling,” using her gestures. There is no mention of any submodality shifts, so this is simple behavioral rehearsal (chaining), without even using the word “swish.” Since the change is contextualized to Sunday evening, it will only occur then, even though it might be useful at other times. (2:46)

16. Keith Livingston describes (rather than demonstrating) how to change an image of a cigarette when they answer the phone that “takes them to a bad place. How would you like to feel when you answer the phone?” (feeling rather than behavior or self-image) “Can you see yourself relaxed and confident?” Then he “swishes” the pictures using size, not mentioning brightness or any other submodality. Keith talks about rehearsing this repeatedly, but with no mention of a break state in between swishes.



All these examples are from trainers who are confident enough about their knowledge and skill to teach others by publicly demonstrating. This makes me wonder how much more variation there is among all the practitioners who didn’t make videos. The kindest thing one could say about all these variations is that they can’t all be correct. These trainers either never read the original sources mentioned at the beginning of this article, or they forgot many important aspects of what they read — or they were taught by someone who didn’t. The wide variation also indicates that most practitioners don’t understand the key principles that underlie each aspect of the pattern.

This brief survey indicates that the problem with NLP’s public “image” is not just a result of a few mavericks who gave the field a bad name. Nor is it just a result of the inappropriate research, the NLP trademark lawsuit, or the Bandler murder trial. The problem goes much deeper than that, to the lack of any kind of quality control over the processes used. It seems likely that a similar review of different people teaching the phobia cure, or any other NLP pattern would show the same kind of essentially random variation. Unless we can come to some kind of agreement about what we do, NLP will continue to be — and be seen as — no different from astrology, numerology, aromatherapy, crystal healing, or all the other bloviations out there.

Of course it’s possible that the original swish pattern can be improved; perhaps one or more of the different ways of doing the pattern is more effective than the original. In that case, anyone proposing changes in the pattern ought to be able to provide a principle for the change, and/or explain how the principles supporting an aspect of the original swish are erroneous. That could be the basis for some interesting discussion that would result in agreement about the best way to do a swish, and perhaps to modifications that would make it work more dependably. That kind of discussion is an essential part of the development of any scientific field, but it is entirely lacking in NLP. If anyone would like to respond to this article, I’m willing to offer you space for it in a future blog post; send your thoughts to me at andreas [at] qwest [dot] net.