7. Submodalities: The Dials Behind the Picture
The moment a memory shifts from a threat to a triviality, the content has not changed; the camera has moved. You are looking at the same face, hearing the same words, feeling the same pressure on your chest, yet the alarm has ceased. This is not a negotiation with the event. It is an adjustment of the submodalities—the fine-grained parameters of the representation that determine whether a sensory code registers as an emergency or as background noise. The nervous system does not respond to the abstract meaning of a symbol; it responds to the format in which that symbol is encoded. When the format is close, loud, and moving, the body prepares for combat. When the format is distant, small, and still, the body permits the integration of experience. Submodality work is the deliberate engineering of that format.
To manipulate format, you must first see the dials. Within the visual channel, the list extends far beyond the clichés of brightness and color. You have location, distance, size, brightness, contrast, saturation, focus, motion, speed, direction, frame quality (continuous versus still), and integration of the visual field. Crucially, and consistently omitted in introductory seminars, are associated versus dissociated viewing, and location within physical space. To view associated is to see through your own eyes, to inhabit the body located at the point of experience. To view dissociated is to step back and see the scene from a third-person perspective, observing yourself and the environment. The second critical parameter is location: is the representation floating in the center of your vision, to the left, to the right, behind your eyes, or projected onto a wall? The auditory channel carries tempo, timbre, volume, pitch, rhythm, distance, source location, and the distinction between internal and external speech. The kinaesthetic channel offers intensity, location, temperature, texture, movement, pressure, and duration. Every thought is a complex mixture of these parameters. Most people operate with a single, unexamined configuration, treating the resulting state as an immutable fact of character rather than a temporary formatting choice.
The leverage of submodality work does not reside in linear adjustments. Turning a dial from 10 to 11 on a volume knob rarely changes the state, nor does shifting a picture two inches to the left. The nervous system operates on threshold effects. There is a digital quality to the shift from safety to threat, from engagement to resistance. The system holds a configuration until a parameter crosses a specific threshold, at which point the state flips abruptly. This is why altering "content" often fails: you are adjusting a parameter that sits deep in the flat region of the response curve. You are turning a dial that the system has already locked. The insight here is that the parameter list is not a menu of equal options; it is a topography of leverage. For any given memory, for any given person, a small number of parameters act as drivers, while the rest are merely decorative. The drivers are the parameters that, when altered, produce the threshold crossing. The work is not to tweak everything; it is to find the few dials that actually move the needle.
You cannot find the drivers by guessing. You must map the contrast. Elicit two memories. The first is a compelling memory: a moment of intense emotion, whether positive or negative, where the representation is vivid and the state is strong. The second is a neutral memory: a moment of indifference, where you recall the event accurately but feel nothing. Guide the person to hold the compelling memory fully, associated, and notice the submodalities. Then, shift to the neutral memory, keeping the content identical in fact—same time, same place, same people—and notice what changed. List every difference. Is the compelling memory larger? Closer? In motion? Associated? Black and white or color? Loud or quiet? Heavy or light? The critical step is to list these differences before changing anything. You are building a map of the current architecture. If you begin by forcing the compelling memory to look like the neutral one, you may miss the driver. The driver is the parameter where the contrast is sharpest and where the state change is most sensitive.
With the contrast list in hand, you proceed to the driver-finding protocol. This requires the baseline established in the previous chapter. You must measure the state change against that baseline, not your own impression of the person. Alter one parameter from the compelling memory to match the neutral memory. If the compelling memory is large, make it small. If it is close, move it back. If it is associated, dissociate it. Pause. Measure the state. Has the emotional charge shifted? Has the physiological response changed? If the state has not shifted, restore the parameter to its original value and move to the next parameter on the list. Repeat this process for each parameter, ranking them by the magnitude of state change they produce. The parameters that cause a significant shift are the drivers. The parameters that produce no change are decorative. You may find that only two or three parameters drive the entire state. In some cases, the driver is not a submodality at all, but a cognitive distortion; the protocol will reveal that as well. The discipline here is to refuse to assume. You let the data tell you what matters.
Once the drivers are identified, you can effect a change. The most rapid procedure is the swish. It is a sequence of operations that rewrites the cue-response chain. Here is the protocol, complete in every step. First, construct the cue picture. This is the image that triggers the unwanted state—the compelling memory, fully associated, large, bright, and vivid. This is the image you want to replace. Second, construct the desired-self picture. This is the image of yourself handling the situation well, or embodying the state you wish to have. It must be small, distant, dark, and dissociated. This is the image that will eventually take over. Place the cue picture in the center of the visual field and the desired-self picture in the lower corner, small and dark. The swish is the rapid movement of the desired-self picture from the corner to the center, expanding to fill the frame, while the cue picture shrinks to a tiny black dot and vanishes. The movement must be fast. The cue picture must go from center to corner/dot, and the desired picture must go from corner/dot to center/expanding. The direction is crucial: the desired image expands and moves toward the viewer, the cue image shrinks and moves away. There must be a blank screen between the movements. After the cue disappears and before the desired image fully forms, a fraction of a second of emptiness must occur. This blank screen breaks the old association. Repeat this sequence five times. On the first pass, you can move the pictures slowly to ensure you have the sequence correct. On the next four passes, move them as fast as you can, creating a whooshing sound and a flash of light. The speed is essential; the system must not have time to construct a counter-argument. Test the change. Present the cue picture. If the state has shifted, the swish worked. If the state remains, the swish failed. The two steps that are usually botched are the construction of the desired-self picture and the timing of the blank screen. If the desired-self picture is not vivid enough, or if the blank screen is too long, the swish will not take. The desired-self picture must be compelling enough to compete with the cue; the blank screen must be a gap, not a pause.
The swish operates within a single memory. When you wish to apply the change to new contexts, you use mapping across. Mapping across is the process of identifying the structural isomorphisms between the original memory and a new situation, and transferring the submodalities from one to the other. You identify the drivers in the original memory. You find the equivalent parameters in the new situation. If the driver is "dissociated," you dissociate the new situation. If the driver is "small," you make the new situation small. You run the swish on the new situation using the same drivers. Mapping across fails whenever the two representations differ in content as well as in coding. If the drivers you identified in the first memory were specific to the content, they will not transfer. The diagnosis is to check the drivers. If the drivers are purely submodal, the mapping will succeed. If the drivers include content-specific elements, you have not found the true drivers. The error is almost always the assumption that "large" is the driver, when "large" is merely a correlate of a specific content structure. The mapping protocol requires you to trust the drivers over the content. If the content differs, the structural transfer is the only valid path.
The mechanism of distancing has genuine support in the literature. Research on imagery vividness shows that the brain processes high-vividness visualizations with greater physiological arousal than low-vividness ones, confirming that format carries signal. Studies by Ethan Kross and Ozge Ayduk on self-distancing demonstrate that referring to oneself in the second or third person reduces emotional reactivity and improves decision quality, a linguistic marker of dissociation that aligns with the submodality shift. Mindfulness trials on decentering show that observing thoughts as mental events rather than facts reduces suffering, which is the functional equivalent of dissociating from the content. These findings support the claim that format manipulation can alter state. The swish protocol itself, however, remains substantially under-tested. There are no rigorous, double-blind trials of the swish sequence. The evidence for the mechanism—threshold effects, cue-dependence, format sensitivity—is robust; the evidence for the specific procedure as a complete intervention is anecdotal. This distinction matters. You are not using the swish because a study proves it works; you are using it because the mechanism is sound, the protocol is precise, and the result is reproducible within the system. The burden of proof lies on the mechanism, not the ritual.
Every framework has an edge. The failure mode of submodality work is distancing a signal that was accurate and should have been listened to. If a person holds a compelling, associated, large, close image of a threat, and you successfully dissociate it, make it small, and move it back, you may have removed the warning system. The distressing picture may not be a distortion; it may be a correct perception of a bad situation. The failure occurs when the practitioner treats the format as the only variable, ignoring the semantic content. If the content is "the car is sliding toward me," and you make the car small and distant, you have made the person stupid, not calm. The constraint is ecology. Before you run the swish or mapping across, you must ask: Is this signal protecting something? Does the person have a viable alternative strategy? If the answer is no, the distress is adaptive. The failure mode is to confuse the removal of pain with the resolution of danger. The work is to distinguish between a signal that is screaming because the format is broken, and a signal that is screaming because the world is breaking. The test is the well-formedness of the outcome: does the change leave the person with a capacity to handle the situation? If the change removes the capacity, it is not a valid change.
The turn this chapter demands is not a technique but a reorganization of the field. We have spent decades arguing over whether NLP is a theory of mind or a set of metaphors. Submodalities dissolve that debate. They show that meaning rides on format at least as much as on content. The same event, coded close and moving, is an emergency. The same event, coded small and still, is a memory. The content is constant; the format determines the state. This means that what a person calls their temperament is often a set of formatting defaults that nobody ever chose. The person who is "anxious" may simply be holding their representations associated, large, close, and moving, without realizing that this is a choice of format, not a fact of nature. The person who is "depressed" may be holding their representations dark, distant, slow, and heavy. The insight is that personality is not a fixed structure; it is a dynamic configuration of submodalities. You can change the configuration, and you change the person. This is not a promise of instant transformation. It is the recognition that the structure is accessible, and that the lever is there. The profound consequence is that no one is trapped in their state. They are trapped in their formatting. And formatting can be edited.
The work does not end with understanding. You must build your own driver list. Select three memories: one compelling, one neutral, and one desired. Elicit the submodalities of each. List the differences. Run the driver-finding protocol on the compelling memory. Identify the two parameters that move you most. Write them down. These are your drivers. Then, select a small unwanted habit—not a trauma, not a crisis, but a habit. Construct the swish using your drivers. Run the swish five times, slowly, then fast, with the blank screen. Test it. If it works, test it twice in real conditions, not in your chair. Go to the situation where the habit occurs. If the state has shifted, you have the change. If not, return to the protocol. Find the driver you missed. The discipline is to trust the data. You are learning to read the structure beneath ordinary speech, to hear the dials behind the picture, and to change them deliberately. That is the skill. That is the distance between a practitioner and a convert. The convert believes the format is the content. The practitioner knows the format is the lever. Use it.
Verification follows immediately, because a shifted state that cannot be sustained under friction is merely a temporary override, not a structural edit. To verify, you introduce the original trigger in low-stakes conditions and measure the latency of the response. A genuine edit collapses the latency curve; the reflex does not fire, or fires in a different register. Consider the 2018 case of a mid-level supply chain director in Rotterdam who automated a chronic latency in vendor negotiations. The protocol identified two primary drivers: spatial compression and tonal resolution. When he rehearsed the swish, he reported an immediate reduction in cortisol spikes, measured via wearable telemetry. Yet three weeks later, the latency returned when a third-party contractor intervened. The mechanism failed because the edit had not been anchored to the hierarchical context of stakeholder authority. The original protocol modified the immediate somatic response but left the secondary representation intact. Structural change requires rewriting the dependency chain, not just the surface signal. You must map the hierarchy of submodalities, identify the governing parameter, and restructure the sequence so the new response becomes the default output rather than an overridden input. The failure mode arrives when the operator treats the protocol as a universal fix rather than a context-sensitive adjustment. Over-application without contextual calibration produces somatic fragmentation; the nervous system decouples from the intended signal, leaving the operator in a state of calibrated confusion, responsive to nothing. The edge is narrow. It requires precise attention to the ecological impact of the edit, not merely its immediate efficacy. Recognizing this edge dissolves the assumption that change is a matter of force or repetition. Change is a matter of alignment. When the structure is correct, the system reorganizes itself without effort. The practitioner stops managing the symptom and begins attending to the architecture. That attention is what separates temporary relief from lasting capacity. You have the tools. You have the protocol. The rest is practice.
Worked Examples — Chapter 7
The difficulty with submodalities is not that the dials are hard to turn. Anyone can ask a person to make a picture dimmer, and almost anyone can do it. The difficulty is that turning a dial produces a number, and a number produced once, in the room, by an interested party, on a person who wants to please them, is very nearly worthless. Everything in this chapter rests on the state measurement you installed in Chapter 6, and the first thing that measurement buys you is not sensitivity. It is the right to ignore most of what you see.
The symbols, once, so that nothing below is ambiguous:
| symbol | meaning |
|---|---|
| S | the composite state-intensity reading from Ch. 6, on the anchored 0–100 scale, higher = more intense |
| S₀ | the baseline reading, the memory untouched |
| Δᵢ | S₀ − Sᵢ, the reduction produced by changing submodality i. Positive Δ means the intensity fell. |
| ε | the within-session measurement noise SD — the same reading, minutes apart, nothing changed |
| ω | the occasion (day-to-day) SD — real variation in the person's state |
| τ | the stable component SD — how charged this memory is for this person, in general |
| r | test–retest correlation between two readings separated by the interval you intend to claim your effect over |
| I | the interaction term, Δ(a,b) − (Δa + Δb) |
| R | the redundancy coefficient, −I / min(Δa, Δb) |
| D | the driver index, Δᵢ / SE(Δ) — an effect measured in units of its own noise |
A note on the clients below. All three are composites, and their numbers are constructed so that the arithmetic is clean. The structure is what transfers; the digits are scaffolding, and you should distrust any teacher — including this one — who hands you a clinical anecdote with four-significant-figure precision and no error bars.