The chart shows a simple line where the shore meets the water. I am looking at it now, spread across the cockpit of my kayak, anchored in a small bay somewhere in the Broughton Archipelago. The line on the chart curves gently—a smooth arc suggesting easy passage around the headland.

But when I look up from the paper, the shore is something else entirely. Rock juts and recedes in patterns within patterns. A cove opens into a smaller cove, which narrows into a channel barely wider than my paddle, which opens again into a pool I could not have seen from twenty meters away. Barnacles roughen the granite at the waterline. Bull kelp sways in water so clear I can see the bottom fracturing into its own complexity—sand, then stone, then the dark mouths of crevices descending beyond sight.

This is a fractal coastline. The closer you look, the more there is to see. Measure it with a kilometer stick and you get one number. Measure it with a meter stick—following every inlet, every cove—and the coastline grows longer. Measure it with a centimeter stick and it grows longer still. There is no true length. The map is not the territory, but the problem runs deeper: there is no single territory to map.

I have been thinking about this lately as I watch the trauma field proliferate its categories. When I began this work four decades ago, we spoke of fight or flight—two responses, clean and memorable. Then freeze was added, and fawn—the appeasement response, the instinct to please or comply to avoid threat. Now I see articles naming flop, flail, faint, and friend. The alphabet grows. Each new letter promises precision, a finer-grained chart of the nervous system’s coastline.

But I have sat with enough people to know that the coastline exceeds every chart we draw of it.

The Trouble with Naming

When we name something, we tend to reify it—to treat it as if it has fixed boundaries and stable properties. We begin to diagnose people as fight types or freeze types, rather than recognizing these as momentary patterns in a constantly shifting system. We develop separate treatments for separate responses, when perhaps what is needed is a more fluid understanding of how nervous systems adapt across time and context.

I do not mean to dismiss the value of these frameworks entirely. Models are useful—indispensable, even—for organizing clinical observations and guiding interventions. I use trauma models extensively in my own work. But they are not the things they describe. The chart is not the archipelago. And the danger lies in mistaking our convenient categorizations for the irreducible complexity of actual human beings responding to threat.

A chart is a functional simplification. It preserves what matters for navigation while deliberately leaving out what would make the map unreadable. Cosmologist Andrew Jaffe makes this point elegantly in The Random Universe: we literally cannot understand anything about the world without models. Not just scientists—everyone. From birth, we build stories about how the world works, test them against experience, and revise them when they fail. Science just does this more systematically.

The key insight is that scientific knowledge is provisional. Always. Newton’s gravitational model worked nearly perfectly for three hundred years. Then Einstein showed it was incomplete. Does that mean Newton was wrong? Not exactly. His model remains brilliantly useful for most applications. We still use it. But it is not the final truth about gravity—there is no final truth. Just increasingly refined models that fit observations better across broader domains.

Consider Jaffe’s example of the London Tube map. Geographically, it is wildly inaccurate. Distances are wrong. Directions are wrong. But functionally? Perfect. It shows you all the connections you need without cluttering your attention with geographical precision that would make navigation harder. The model preserves what matters for its purpose while deliberately simplifying everything else.

That is what good clinical models do. They are functional simplifications. The trouble starts when we forget they are simplifications.

The Turn Toward Complexity

What strikes me is that while trauma theory has been proliferating categories—adding new letters to the alphabet—much of contemporary science has been moving in precisely the opposite direction. Toward understanding biological and psychological phenomena as complex dynamical systems that resist simple categorization.

In neuroscience, dynamical systems theory has become increasingly influential for understanding how brain networks coordinate and adapt over time. Rather than treating neural activity as a series of discrete states, this approach recognizes that the brain exists as a continuous, nonlinear system where patterns emerge, dissolve, and reorganize. As one researcher notes, studying systems like biological organisms and their behavior may require an understanding of interactivity, dynamics, and emergent order rather than fixed categories.

This shift extends to psychotherapy research. A systematic review found growing application of dynamical systems and complexity theory to understand how change unfolds in therapeutic relationships—methods that account for nonlinearity, phase transitions, and self-organization. Concepts that stand in stark contrast to categorical diagnoses and treatment protocols.

Perhaps most relevant to our understanding of trauma, research on physiological regulation has revealed profound fractal organization across multiple systems. The healthy human heartbeat, far from being metronomically regular, displays complex fractal variability—self-similar patterns repeating at multiple time scales. Remarkably, this complexity diminishes in illness. Patients with congestive heart failure show less variable, more predictably regular heart rhythms. Classical homeostatic models assumed the body maintains stability through constancy, returning to a set point after each perturbation. But these findings suggest something different: health requires complexity, not constancy. The capacity to vary, to respond flexibly, to never quite repeat—this is what a well-functioning system looks like.

Neuroscientist Gerd Werner documents fractal organization at all levels of the nervous system: from the branching patterns of individual neurons to the temporal dynamics of brain activity. The nervous system, like trees and coastlines and river deltas, displays self-similarity—patterns that repeat at different scales. A neuron’s dendritic tree branches like a river delta, which branches like a circulatory system, which branches like neural networks. The same organizational principles operate across vastly different scales.

Sitting in my kayak, watching the shoreline fracture into smaller and smaller detail, I recognize this. The inlet mirrors the archipelago. The crevice mirrors the inlet. The pattern repeats, all the way down.

The Problem of Categories

Asking “How many trauma responses are there?” is like asking how many inlets exist along this coast. The answer depends entirely on the scale of measurement and where you draw your boundaries.

Consider what self-similarity means in practice. A small section of a fractal object, when magnified, resembles the whole. The edge of a coastline shows the same jagged complexity whether you are viewing it from a satellite or examining a single meter of shoreline. Similarly, trauma responses may show self-similar patterns at different scales of observation.

At the level of immediate physiology: heart rate changes, muscle tension, breathing patterns. At the level of behavioral response: fight, flight, freeze, or various combinations. At the level of sustained adaptation: anxiety, depression, dissociation. At the level of life patterns: relationship styles, career choices, addictive behaviors.

Each level displays recognizable patterns that echo patterns at other levels. Someone who freezes in immediate threat may also freeze in their life—withdrawing from engagement, waiting passively for change that never comes. Someone who fights in crisis may also maintain a fighting stance in their approach to work, relationships, and self-care.

The developmental psychologist Esther Thelen described this beautifully in her work on motor development: what appears to be a distinct developmental stage is actually a temporary stable pattern in a continuously changing system. The pattern emerges, holds for a while, then dissolves as conditions change.

Perhaps what we call fight, flight, and freeze are not three distinct programs but rather temporary stable patterns—what dynamical systems theorists call attractors—that emerge from the interaction of multiple factors: threat intensity, available energy, prior learning, social context, and countless others. In some conditions, the system settles into fight. In others, it tips into flight. In still others, freeze becomes the dominant pattern. But these are descriptions of moments, not fixed types.

What I Actually Do in the Room

When I am trying to determine whether someone is displaying a fight response or a fawn response or a flop response, I am not actually present to the dynamic, shifting quality of their experience. I am sorting them into boxes.

But when I simply follow the patterns without needing to name them precisely—when I notice how energy moves in their system, what happens when safety increases or decreases, how their responses shift across contexts—something different becomes possible. I am tracking a process rather than applying a diagnosis. I am curious rather than certain.

This does not mean abandoning frameworks entirely. The language of fight, flight, and freeze can be extremely useful, particularly when it helps someone recognize their own patterns and understand them as adaptive rather than pathological. Van der Kolk’s work has been transformative in helping trauma survivors understand that their symptoms make sense—that they are not broken but responding logically to overwhelming experiences. Herman’s framework for understanding complex trauma provided essential validation for survivors whose experiences did not fit neat diagnostic categories.

But there is a difference between using a model as a heuristic—a way of organizing observations—and believing the model captures the actual structure of the phenomenon. When we assert that specific symptoms always indicate specific types of trauma, or that specific treatments match specific response patterns, we may be imposing order that does not actually exist in the system.

I think of this when I paddle these waters. The chart helps me navigate. But if I trusted only the chart, I would miss the hidden channels, the unexpected passages, the way the tide transforms what is possible. The chart is where I start. It is not where I stay.

A Cautionary Example

One particularly vivid illustration of this tension appears in Polyvagal Theory, which has become extraordinarily influential in trauma treatment. The theory proposes three distinct neural circuits corresponding to social engagement, fight/flight, and shutdown—a neurobiological threefold categorization that maps neatly onto clinical observations.

The appeal is obvious: it provides a clear, biological explanation for phenomena therapists observe daily. It suggests specific interventions. If someone is in dorsal vagal shutdown—a state of collapse and disconnection—we help them climb back up the autonomic ladder. If they are in sympathetic activation—the mobilized, hypervigilant state—we help them access ventral vagal regulation and social engagement.

Yet the theory has also faced substantive criticism. A systematic review found limited evidence supporting the predicted relationships between vagal tone and social behavior. The neat three-circuit model may oversimplify neural regulation that is actually far more distributed and interactive. The proposed evolutionary sequence—where newer circuits supposedly inhibit older ones—does not align well with current understanding of how evolution actually works.

This does not make Polyvagal Theory useless. Many clinicians find it helpful for organizing interventions. I have used it myself. But it illustrates how our desire for clear, biological explanations can lead us to embrace models more strongly than the evidence warrants.

The trajectory of Polyvagal Theory is typical of all models: useful, but not precisely fitted to the reality of human beings. As Jaffe notes about cosmological models, we do not get to certainty. We get to “this model fits the current observations better than alternative models, so we will use it until something better comes along.”

That is not cynicism. It is epistemic honesty.

Working With Complexity

What might it mean to work with complexity rather than trying to reduce it? Several principles emerge from sitting with this question—and from sitting with people whose nervous systems refuse to match my charts.

First, I track patterns rather than labeling types. Instead of asking “Is this person a freeze type or a fawn type?” I ask “What patterns am I noticing in how this person’s system responds to different conditions?” The shift is subtle but significant. I am looking at process rather than imposing categories.

This is not to say that patterns are meaningless or that everyone shifts constantly. Some people do show remarkably consistent response tendencies—a person who reliably freezes under stress, or whose system defaults to appeasement in the face of conflict. These tendencies are real and clinically meaningful. The point is not to deny them but to hold them as descriptions of what a nervous system tends to do, not fixed diagnoses of what a person fundamentally is. The pattern is what happens. It is not the whole of the person.

Healthy systems show complex variability, not predictable regularity. Someone who always responds the same way to threat may actually be stuck in a rigid pattern—a sign of dysregulation, not stability. Flexibility—the capacity to shift responses as context changes—may be a better indicator of healing than consistent adherence to any single pattern.

Whether it is Polyvagal Theory, attachment styles, or the ever-expanding alphabet of trauma responses, I hold models lightly. Models help us organize observations and guide interventions. But they are tools, not truth. The moment we mistake the model for the actual complexity of the person in front of us, we have stopped being truly present.

Some aspects of human experience simply cannot be reduced to models. The grief of the parent who lost a child, the rage of the person whose trust was violated, the despair of someone who sees no way forward—these are not symptoms to be categorized. They are human responses that deserve to be witnessed in their full, terrible specificity.

Recovery as Fractal Process

If trauma responses show fractal, self-similar organization, what might this suggest about recovery?

Traditional treatment models often present healing as a linear progression: from acute symptoms to reduced symptoms to eventual resolution. But fractals are not linear. They are recursive, with patterns repeating at different scales. Recovery works this way too.

Someone might resolve acute PTSD symptoms—the flashbacks and hypervigilance diminish, sleep improves, they can talk about what happened without dissociating. Real progress. Yet years later, in a different context, they might encounter an experience that activates a familiar pattern at a different scale. Not a failure of treatment but rather the self-similar quality of nervous system organization revealing itself.

This suggests that recovery is not about permanently eliminating patterns but rather about increasing flexibility—the capacity to move through patterns without getting stuck, to recognize them without being controlled by them. It is about adding complexity rather than achieving some imagined perfect stability.

Peter Levine describes this beautifully in his Somatic Experiencing work—a body-based approach that attends to the physical sensations and movements the nervous system generates in response to threat. Healing happens not by forcing change but by creating conditions where the nervous system can complete cycles that trauma interrupted. We are not fixing broken people. We are facilitating natural processes that got stuck. The nervous system already knows how to heal; our job is to stop interfering with that knowing.

And yet—the nervous system does not heal alone. Trauma so often occurs in relationship: the violation of trust, the failure of protection, the absence of witness. Healing, too, requires relationship. The co-regulation that happens when one nervous system attunes to another. The experience of being seen without being fixed. The slow, iterative repair of the capacity to trust. These relational dimensions do not fit neatly into our models of individual nervous system function, but they may be where the deepest healing happens. The coastline is not just the rock meeting the water. It is shaped by the tides.

I think of the tide moving through these islands. It does not flow in a straight line. It finds the channels, the hidden passages, the routes that open and close with the hours. Recovery moves the same way—not linearly but through whatever passages become available, often surprising us with where it emerges.

The Humility of Uncertainty

After four decades of clinical work, I am less certain about trauma responses than I was when I began. I have watched too many people defy categories, recover through means that were not supposed to work, fail treatments that should have succeeded. I have witnessed healing emerge from relationships that had nothing to do with trauma expertise, and I have seen technically excellent trauma treatment accomplish nothing.

This uncertainty is not resignation. It is recognition that human beings are genuinely complex—not just complicated in ways we have not figured out yet, but complex in ways that resist complete mapping. Models help us navigate this complexity, but they are provisional tools, not final answers.

The same holds for trauma work. Our models fit many observations. They guide useful interventions. But they are not the final word on how nervous systems respond to threat and organize recovery. They are our current best guesses, refined through decades of clinical observation and neuroscience research, but still provisional.

Not everything that matters can be measured. Not everything that helps can be manualized. And not every person will fit the model, no matter how comprehensive we make it.

That provisionality, properly understood, is liberating. It means we can use models without being enslaved to them. We can appreciate what they illuminate while remaining attentive to what they obscure. We can hold our frameworks lightly enough to notice when the person in front of us is showing us something our model does not account for—and trust that observation more than our theory.

The tide is turning. I fold the chart and tuck it into the dry bag, then lift my paddle from where it rests across the cockpit. The water moves beneath me, finding its way through channels the chart does not show.

Somewhere ahead, beyond the headland, there are passages I have not paddled before. The chart suggests one route. The water will show me others. Both are true. Both are incomplete.

I dip my paddle and pull toward the open water, toward whatever the coastline reveals when I arrive close enough to see it.

Questions for Reflection

  • What patterns do you notice in your own responses to threat or stress? Do these patterns shift across different contexts, or do you tend toward one primary response?
  • When you think about your own healing or growth, does it feel more like a linear progression or like a process with repeating themes at different scales?
  • How might it change your clinical work—or your understanding of your own experience—to think of trauma responses as temporary attractors in a complex system rather than as fixed types?
  • What would it mean to hold your clinical models—or your understanding of yourself—more provisionally?

Sources

Beer, R. D.(2014). Dynamical systems and embedded cognition. In K. Frankish & W. M. Ramsey (Eds.), The Cambridge handbook of artificial intelligence (pp. 128-150). Cambridge University Press.

Favela, L. H.(2020). Cognitive science as complexity science. Wiley Interdisciplinary Reviews: Cognitive Science, 11(6), e1525. https://doi.org/10.1002/wcs.1525

Goldberger, A. L., Peng, C. K., & Lipsitz, L. A.(2002). What is physiologic complexity and how does it change with aging and disease? Neurobiology of Aging, 23(1), 23-26.

Haeyen, S. (2024). Polyvagal theory and art therapy: Theoretical and clinical perspectives. Frontiers in Psychology, 15, 1352025.

Herman, J. (1992). Trauma and recovery: The aftermath of violence—from domestic abuse to political terror. Basic Books.

Herman, J. (2023). Truth and repair: How trauma survivors envision justice. Basic Books.

Ivanov, P. C., Amaral, L. A., Goldberger, A. L., Havlin, S., Rosenblum, M. G., Struzik, Z. R., & Stanley, H. E.(1999). Multifractality in human heartbeat dynamics. Nature, 399(6735), 461-465.

Jaffe, A. H.(2025). The Random Universe: How Models and Probability Help Us Make Sense of the Cosmos. Yale University Press.

Klocek, A., Premus, J., & Řiháček, T. (2023). Applying dynamic systems theory and complexity theory methods in psychotherapy research: A systematic literature review. Psychotherapy Research, 34(6), 828-844. https://doi.org/10.1080/10503307.2023.2252169

Levine, P. A.(2010). In an unspoken voice: How the body releases trauma and restores goodness. North Atlantic Books.

Porges, S. W.(2021). Polyvagal theory: A biobehavioral journey to sociality. Comprehensive Psychoneuroendocrinology, 7, 100069. https://doi.org/10.1016/j.cpnec.2021.100069

Porges, S. W.(2023). Our polyvagal world: How safety and trauma change us. W. W. Norton.

Seguin, C., Sporns, O., Zalesky, A., & Calamante, F. (2023). Network communication models improve the behavioral and functional predictive utility of the human structural connectome. Network Neuroscience, 7(2), 661-684.

Thelen, E., & Smith, L. B.(1994). A dynamic systems approach to the development of cognition and action. MIT Press.

Van der Kolk, B. A.(2014). The body keeps the score: Brain, mind, and body in the healing of trauma. Penguin Books.

Werner, G. (2010). Fractals in the nervous system: Conceptual implications for theoretical neuroscience. Frontiers in Physiology, 1, 15. https://doi.org/10.3389/fphys.2010.00015