Why Writing the Body Works (and Why It Sometimes Doesn't)

What Pennebaker and his collaborator Joshua Smyth do over the long arc of their research is push past the whether-it-works question into the how-it-works question. Several mechanisms are now reasonably well-established. Writing in a narrative form — with a beginning, middle, and end — outperforms writing the same emotional content as a fragmented list, suggesting that part of what is doing the work is the construction of a coherent story rather than the discharge of feeling.[^4] Writers whose health improves the most are not those who use the most negative-emotion words but those whose use of cognitive words (because, realize, understand, cause) increases over the four days, suggesting that the meaningful change is the writer's gradual movement from inhabiting the experience to making sense of it. Working memory improves after expressive writing — students who write about coming to college get higher grades the following year — which fits a general picture of the writing as freeing up cognitive resources that had been bound up in not-thinking-about a difficult thing.

What does not seem to do the work: catharsis as such. Simply venting, in the raw sense, has a poor track record. The writing that helps is the writing that, somewhere across its four days, begins to organize what it is rendering. This is not a moralistic point about effort. It is a description of what the data show. Words like because and realize and the reason that show up more often, on day four, in the essays of the people who later get sick less.

II. The somatic vocabulary

If Pennebaker is the protocol, Peter Levine is the phenomenology. Levine, a behavioral physiologist by training, spent decades watching the way mammals — particularly large prey animals — recover from near-death encounters in the wild. The central observation is one that has by now been absorbed into so many therapeutic schools that its origin is sometimes lost: that animals who survive a predator encounter do something that human beings under threat often fail to do. They shake. They tremble. They cycle through a few minutes of involuntary movement, and then they get up and walk on, and they do not appear to develop the lingering symptoms of what we call traumatic stress.

Levine's argument, worked out in Waking the Tiger and the long therapeutic tradition that became Somatic Experiencing, is that the immobility response — the freeze that overtakes a body when neither fight nor flight is possible — is not in itself the problem. The problem is when the immobility persists, when the energy mobilized for survival never gets discharged, when the nervous system stays in a kind of arrested high-arousal state long after the original threat is gone. As he puts it:

Traumatic symptoms are not caused by the "triggering" event itself. They stem from the frozen residue of energy that has not been resolved and discharged; this residue remains trapped in the nervous system where it can wreak havoc on our bodies and spirits.[^5]

What follows from this physiological reading is a working vocabulary that has, in the last decade, been borrowed by every adjacent discipline, often without much of its original precision. Three terms are doing most of the work.

Felt sense is borrowed from the philosopher and psychologist Eugene Gendlin and means something specific. It is the bodily quality of an experience prior to its being named — the heaviness in the chest, the tightening at the throat, the not-quite-locatable quality of something is wrong here that announces itself before the cognitive labeling kicks in. Levine treats the felt sense as the language of the older parts of the brain. To work somatically is to attend to it before, or alongside, the verbal. Sensation is the language of the reptilian brain, in Levine's formulation. A person who has lost contact with their own felt sense — and who can, under questioning, only report what they think about a given subject, never what they feel toward it — has, in a real sense, become a stranger to the part of themselves that holds the original imprint.[^6]

Titration is the term Levine borrowed from chemistry, where it names the slow drop-by-drop addition of one solution to another to allow controlled reaction. In the somatic-trauma context, titration names the deliberate practice of approaching a difficult memory or sensation in small enough doses that the nervous system is not overwhelmed. The opposite of titration is what older trauma traditions called abreaction — the attempt to "break through" by reliving the event in its full intensity, trusting that catharsis will resolve what restraint would not. Almost all modern somatic work treats abreaction as having been a long professional mistake. The doses are small on purpose. A person who has just touched the edge of a difficult sensation is asked to come away from it before going further.

Pendulation is the partner concept. The nervous system, Levine argues, has its own rhythm of contraction and expansion; it does not move directly from disturbance to resolution but oscillates between them. Pendulation names the deliberate cultivation of that oscillation — touching the difficult thing, then deliberately turning attention to a place in the body that is not in distress, then back to the difficult thing, then back out. The metaphor he uses is the swinging of a pendulum: the measure of the swing one way is the measure of the swing the other. The work is not to push through. The work is to let the system move at its own pace, in both directions, until the original arrest gradually loosens.[^7]

What this means concretely, and where the writing-as-therapy discourse can pick up something genuinely useful, is that the prose of a serious essay about a difficult experience should not be expected to move only in one direction. A piece of writing that drives steadily into the most painful material and stays there is, in Levine's terms, a piece of writing that has lost the pendulation. The writer is being asked to do, in language, the thing that the nervous system was unable to do at the time of the original event: to register the difficulty, and to come back out, and to register again, and to come back out, until what was static begins to move.

This is something good memoirists have always known without naming. The classic memoir paragraph that takes the reader into the worst of it — and then breaks line, and gives a detail of the kitchen, the temperature of the room, the color of the wallpaper — is not a failure of nerve. It is pendulation as craft. The reader, like the nervous system, can only hold so much of the difficult material at once before something has to soften. The writers who can hold this rhythm produce prose the body can stay with. The writers who cannot tend to produce prose that the reader admires from a distance and does not return to.

Two cautions about the somatic vocabulary, though, before going on. The first is that Levine's clinical claims have a much weaker evidence base than Pennebaker's experimental ones. Waking the Tiger is a brilliant work of phenomenology, drawing on careful observation, ethological literature, and a long career of clinical practice. It is not, in the sense the expressive-writing literature is, a body of randomized controlled trials. Somatic Experiencing as a method has accumulated a small but growing empirical literature, with some encouraging results, but the field is younger and the evidence thinner than the cultural enthusiasm sometimes implies.[^8] The vocabulary is useful. The mechanisms it describes are well worth attending to. The strength of the empirical support is not the same as Pennebaker's, and a careful reader should hold the two distinctions clearly.

The second caution is that felt sense, titration, pendulation have, in the last few years, joined the inventory of words that get tossed around by anyone wanting to imply somatic literacy. They are real words. They name real things. They are not, in themselves, evidence that the speaker is doing the work the words describe. As with all professionalized vocabularies that escape into the broader culture, the test of whether the vocabulary is being used with precision is whether the speaker can describe what the word names without the word. A writer who uses titration in three consecutive sentences without slowing down to render the actual approach-and-withdraw of attention is, almost certainly, using the word as decoration.

III. The clinical frame

This brings us to van der Kolk, and to a complication. The Body Keeps the Score is, in its substance, the most important book of the three for the popular discourse, and the one whose place in this discussion is most awkward.

The book itself is a synthesis of forty years of clinical work with severely traumatized patients — combat veterans, survivors of childhood sexual abuse, victims of catastrophic accidents and assaults — and it makes a sustained, evidence-rich argument that the long-term consequences of severe trauma are not primarily a story about distorted cognitions or maladaptive beliefs. They are a story about a nervous system that, having absorbed an event the rational brain could not metabolize, keeps reproducing the conditions of the event at the level of the body itself. The flashback is not a memory in the ordinary sense; it is the body re-running the original alarm. The hypervigilance is not a chosen attitude; it is a physiological state. The dissociation is not a personality trait; it is a survival mechanism that, having been useful once, is now firing in conditions where it is not.

The methodological point van der Kolk presses is that any treatment that addresses the cognitive layer alone — the story the patient tells about the event — without engaging the body that is still living inside the event will, for the most severely affected patients, fail. The book's argument for yoga, for EMDR, for neurofeedback, for theater, for Pilates, for whatever modality engages the body directly, is grounded in this observation: that the patient is not, in the most ordinary sense, "thinking wrong" about the trauma. The patient is in the trauma, somatically, every time the alarm fires.

Van der Kolk's case for writing within this frame is substantial but qualified. He cites Pennebaker approvingly. He describes patients for whom journaling and letter-writing have done real work. He attends, with the precision of someone who has watched it happen, to the way the handwriting of a severely dissociated patient changes when an earlier developmental state takes over the page — to the moments when an adult patient writing about her childhood begins, mid-paragraph, to write in a child's hand. None of this is incompatible with what Pennebaker and Levine describe; it extends what they describe into a more clinically extreme territory.

But where van der Kolk is most useful for our purposes is in his framing of what writing cannot do, and where attempting it without other supports is a mistake. The most-quoted chapter of his book, on Broca's area going dark during a flashback, makes the point at the level of the brain itself:

Deactivation of the left hemisphere has a direct impact on the capacity to organize experience into logical sequences and to translate our shifting feelings and perceptions into words. (Broca's area, which blacks out during flashbacks, is on the left side.) Without sequencing we can't identify cause and effect, grasp the long-term effects of our actions, or create coherent plans for the future. People who are very upset sometimes say they are "losing their minds." In technical terms they are experiencing the loss of executive functioning.[^9]

The writer in the middle of an active flashback is not in a position to write about the flashback. The language faculty itself goes offline during the worst moments of re-experiencing. This is a technical fact about how trauma operates in the brain, and it explains something the expressive-writing literature is sometimes vague about: the patients for whom expressive writing helps the most are patients who have already done some prior work to be able to write about the difficult material in the first place. They are not patients in the middle of an unprocessed flashback. They are patients who have, by some combination of clinical work, time, and circumstance, gained enough distance from the event to be able to put it into language at all.

This is the load-bearing distinction. Pennebaker's protocol is for people who can write about the thing. For people who cannot — and severe-PTSD populations often cannot, for the brain-level reasons van der Kolk lays out — the writing prompt itself can be destabilizing. Not therapeutic; destabilizing. Pennebaker himself notes, in his 2016 revision, that expressive writing as part of a clinical intervention requires the therapist to prepare the client for an immediate spike in distress and to have tools in place for managing it. He goes further: clients with severe ongoing trauma symptoms may benefit from receiving feedback on what they have written; others may benefit from keeping it private. The flexibility is not optional. It is part of how the protocol gets used responsibly.[^10]

This is also where the field publicly disagrees with itself. There is a real and ongoing professional argument about whether the somatic-and-expressive turn in trauma treatment has come at the cost of what was working in older talk-therapy modalities, particularly for patients whose trauma is not catastrophic but cumulative — the small repeated injuries of difficult childhoods, ordinary disappointments, the corrosive effect of prolonged stress that was not, in any individual moment, traumatic enough to register as such. Some of the more recent literature on expressive writing for PTSD specifically has been less encouraging than the early studies, with effect sizes shrinking under more rigorous designs.[^11]

It is also worth noting, briefly, that Bessel van der Kolk has been the subject of professional disputes in recent years connected to his role at the Trauma Center he founded, and that some of those disputes have been bitter. None of that is the subject of this essay, and none of it changes what the book actually argues. The argument is what survives. The book remains the synthesis of an extraordinary clinical career, and the case it makes for the body's place in trauma persists whether or not the institution that produced it does.

IV. What the popular reception got wrong

If this is what the three books actually say, what has the popular reception of them turned them into?

A few things, none of them entirely the books' fault.

The most consequential is the more is better error. The Pennebaker effect sizes are small and time-limited; the popular reception treats expressive writing as if it were a continuous practice that delivers continuous returns. Journal every day for life is not what the research supports. Three or four sessions, fifteen minutes each, when something difficult is asking to be metabolized, with breaks of months or years between rounds as new material accumulates is much closer to what it supports. The journal that fills six volumes a year is, as often as not, a mechanism for keeping a difficult thing chronically warm rather than a mechanism for letting it cool.

The second error is conflating writing with processing. The Pennebaker results are not about producing prose. They are about a specific four-day exercise in a specific format with a specific kind of internal honesty. The diary-keeping that, in his own data, shows up as not particularly health-promoting — most diary entries don't grapple with fundamental psychological issues, in his summary phrase — is, in popular parlance, often what people mean when they say journaling. The expressive-writing protocol and the daily diary look superficially similar. The data treat them as two different things.

The third error is the one most worth naming: the popular reception has tended to assume that what writing is doing for the writer is catharsis — the discharge of pent-up feeling, the relief of having said the thing — and that the more intensely the feeling is expressed, the better the writing is working. This is roughly the opposite of what the data show. The writers who improve the most are the ones who, across four days, gradually move from feeling about the experience to making sense of it. The catharsis happens, but it is incidental. The mechanism is closer to narrative integration than to emotional release. A writer producing maximally raw, maximally intense, maximally feeling-saturated prose on day four is, in Pennebaker's data, no more likely to show health benefits than a writer producing a flat list of facts. The writer producing a coherent story, with cause and effect and because and I think the reason was, is.

The fourth error, more diffuse and harder to argue, is the implicit promise that has accreted around the somatic-trauma literature: that there is, somewhere within reach, a sufficient practice — yoga, journaling, breathwork, somatic experiencing, EMDR — that will resolve the original injury. The literature itself does not promise this. Waking the Tiger is careful to distinguish between symptom resolution and full transformation, and to note that even the latter is a process that may take years. The Body Keeps the Score presents its inventory of modalities as exactly that — an inventory — and is careful to note that no single modality works for everyone, that combinations are usually necessary, and that for many patients the appropriate goal is not resolution but a workable equilibrium. The popular reception has often skipped the qualifications and gone directly to the modality.

V. One specific thing a careful reader can actually try

Because this essay is methodological rather than therapeutic, the most concrete thing it can offer is a faithful description of the Pennebaker protocol, with the modifications that thirty years of subsequent work suggest are reasonable. A reader who wants to try it can do so without a clinician, with the following caveats.

It is not for material that is currently destabilizing you. If the experience you would write about is one whose ordinary recall already produces flashback, dissociation, panic, or a level of distress you cannot easily come back from, this is not the right tool. The first thing to find is a clinician.

It is not a substitute for therapy if you are in therapy, or if you have reason to be. It is, at most, an adjunct.

With those said:

For four consecutive days, set aside fifteen to twenty minutes when you will not be interrupted. Use a place where you will not be observed. Pen and paper are slightly preferred to a screen for the simple reason that handwriting is harder to edit in real time and tends to slow the writer down, but typing is fine.

Write about an experience that has affected you deeply, that you have not talked about in detail with others, and that continues to take some weight in your life. You can write about the same experience all four days or different experiences each day. You can return to the same material as often as you want.

Write continuously. If you run out of things to say, repeat what you have already written until something new comes. Do not go back and edit. Spelling and grammar are not the point. Coherence in the conventional sense is not the point.

Write what happened, how you felt about it, and how you feel about it now. The doubling — the past feeling and the present orientation — is the part that does the work. A piece of writing that stays only in past tense is missing one of the three required pieces.

Across the four days, do not police what you say toward more or less intensity. Let the prose move at the pace it wants to. Some sessions will be flat; some will be heavy; some will surprise you with the appearance of an event you had not expected to write about. All of these are part of the protocol working as designed.

After each session, expect to feel slightly worse for an hour or two. This is consistent with the data: the writing produces an immediate spike in distress that subsides within a few hours. If the spike persists into the next day, or if it is severe enough to interfere with ordinary functioning, stop the protocol and consult a clinician.

When the four days are over, the writing is finished. You do not need to keep it. You do not need to share it. You do not need to do anything with it. If you want to keep it, keep it private. The Pennebaker literature is consistent on this point: writing intended for an audience tends to do less of the work than writing intended only for the writer, because the writer's attention shifts toward managing how it will sound rather than toward what is actually there to say.

After the four days, do not start over the next week. The data do not support a continuous practice. If, in three or six months or a year, more material has accumulated and is asking to be metabolized, do another round. In the meantime, the protocol's work is done.

This is a small intervention. It does not promise much. What it promises, the data largely deliver: a small but real improvement in mood and short-term physical-health markers, a measurable improvement in working memory, a higher likelihood that the writer can think about the difficult material without it commandeering disproportionate cognitive resources. None of this is transformation. All of it is worth fifteen minutes a day for four days.

VI. Where the discourse is going wrong

A last note, before the genealogical retrospective.

The risk in the current moment is not that the writing-the-body discourse is too widely available. The risk is that it has begun to medicalize what should not be medicalized. Trauma has become the word a generation of intelligent, ordinarily unhappy people use to describe ordinary unhappiness. The body keeps the score has become the explanation for difficulties whose actual sources are in the present rather than the past. The protocols developed for, and tested on, populations with serious clinical histories have been extended into the wellness market as universal tools for what would once have been called the difficulty of being a person.

This is not, in itself, a problem with the underlying books. None of the three authors writes as if his protocol were the answer to ordinary unhappiness. Pennebaker is consistent that the expressive-writing effect is small and time-limited. Levine writes for clinicians and for adults working through severe historical material. Van der Kolk's case studies are uniformly drawn from clinical populations. The misappropriation has happened downstream, in the cultural translation.

The case worth holding open is that some unhappiness is genuinely traumatic in the technical sense, and writing the body has a real role in addressing it; and most unhappiness is not traumatic in that sense, and the writing protocols, applied to it, do something different than what they were designed to do. They become, in the second case, a sustained practice of self-attention. There is nothing wrong with sustained practices of self-attention; the pre-modern devotional and contemplative traditions are full of them. But they are not what the trauma literature is talking about, and conflating them obscures what each is actually for.

The honest summary: writing about a difficult experience, in the specific format the research describes, helps a little, for a while, with a particular kind of metabolic work. It is not therapy. It is not transformation. It is a usefully designed brief intervention with a small but real effect size. Treated as that, it earns its place in the working repertoire of an adult life. Treated as more, it begins to do the very thing that the somatic-trauma literature was developed to push back against: a reduction of the body to what can be managed by the conscious mind.

The most useful sentence about the body, in the end, may not be that the body keeps the score. It may be the older, less marketable version: that the body is not, and never was, separable from the person who lives in it, and that the practices we develop for paying attention to it are part of how we become, over time, the kind of person we want to be. Writing — when it is honest, and unhurried, and addressed to the writer rather than to an audience — can be one of those practices. It is not the only one. It is not even, for most people, the central one. It is one good tool among several, used at the right time, in the right format, for the right material.

For some readers, that will be enough. For others, what will be required is the other thing — the room with the clinician, the modality the notebook cannot reach, the slow reconstruction of a relationship to a body that has, for too long, been somewhere the conscious self refused to live. The notebook does not replace any of that. It does, sometimes, get a person to the place where they can begin.

VII. Notes from the genealogical probe

The Foucauldian probe (docs/magazine/FOUCAULDIAN-PROBE.md) is designed for pieces tagged as Coverage Gap, Mosaic, Atlas Briefing, or Constellation. This piece is technically a methodological essay and not formally subject to the modal, but several of the probe's questions surface useful tensions worth naming explicitly before publication.

The minor archive (Q1). The piece works almost entirely from three canonical books — the load-bearing texts of the modern trauma-and-writing discourse. The minor archive here would be the considerable critical literature on the writing-as-therapy movement (Suzette Henke's Shattered Subjects, Nancy K. Miller's writing on autobiographical practice, the more skeptical recent work by Will Storr and others on trauma's cultural inflation), and the patient-side memoirs that have pushed back on the "do the work" register from the inside. The piece names this risk obliquely (the more is better error, the medicalization argument in §VI) but does not cite the minor archive directly. A future revision could cite Storr or Watters more explicitly.

Opening with a particular case (Q2). The piece opens not with a single dated scene but with the cultural claim itself — the body keeps the score — and uses the genealogy of the phrase as the entry. This is closer to opening in synthesis than in case, and it is a deliberate choice for a methodological essay; the price paid is that the reader does not encounter a specific person doing the protocol on a specific day. A possible companion piece could open with one of Pennebaker's actual student writers in 1986, or with one of Levine's case studies (Marius, Margaret), to anchor the methodology in a body.

Discontinuity (Q3). The piece names the rupture between what the underlying books say and what the popular reception turned them into (§IV), which is the most important discontinuity to mark. A discontinuity it does not fully mark is the older history of writing as cure — the long lineage of confessional, devotional, and epistolary practice from Augustine through the Puritan diary to Romantic introspection to mid-twentieth-century psychoanalytic transcription. Pennebaker's protocol is treated as a 1986 invention; in fact it sits inside a much older tradition that the book itself sometimes nods to and the essay does not. Worth a future Mosaic-adjacent companion.

Gap typing (Q4). Not formally a coverage-gap piece, but the implicit gap the essay is filling — the corpus has the trauma-writing books but has not had them placed against each other carefully enough — is discursive rather than acquisitional. Vela has the books; what it has lacked is the synthesis. This essay is an attempt at the synthesis.

What the metric does not see (Q5). Recall@K against the corpus would have surfaced van der Kolk easily (825 passages), Pennebaker adequately (171), and Levine modestly (160). What recall does not see is which of the three authors gets the editorial weight in a given essay. A naive synthesis would have given van der Kolk the spine because his volume is largest; this piece deliberately gives Pennebaker the spine because his protocol is the actionable methodology a reader can use, and treats van der Kolk as framing rather than as primary. The decision is invisible to the metric and load-bearing for the piece.

Reversal (Q6). The most important reversal in the piece is the one in §IV: where the popular discourse assumes more writing is better, the data suggest less writing, specifically dosed, is better; where the popular discourse assumes catharsis is the mechanism, the data suggest narrative integration is the mechanism. A second reversal is the one in §VI: where the cultural register treats the somatic-trauma literature as a tool for self-improvement, the literature itself was developed in opposition to a self-improvement frame — to push back against the implicit claim that the conscious mind can manage what the body has registered. The medicalization the essay names is a return, in different vocabulary, of the cognitive-managerial register the literature was developed against.

The piece thus passes Q3, Q5, Q6 without much qualification; passes Q1 and Q4 with modest acknowledged limitations; and consciously declines Q2's strict template in favor of the cultural-claim opening, which is the appropriate move for a methodological essay aimed at a bridge audience.