Walter Paganin
Rome, Italy

At Marathon in 490 BCE, according to Herodotus, an Athenian soldier named Epizelus was fighting bravely when he suddenly lost sight in both eyes, though no weapon had touched him.¹ He remained blind for the rest of his life. He described having seen an enormous armored figure whose beard covered his shield; the apparition passed him by and killed the man at his side.²

A modern clinician reading this account will recognize a configuration of symptoms: an overwhelming traumatic exposure, a sensory loss without organic basis, and a persistent course. Across the full length of the descriptive record, such configurations recur with a regularity that our modern categories do not quite accommodate.² The question worth asking is not Epizelus’ diagnosis, but where his symptoms would fall in a modern nosology.

Under current criteria, Epizelus would not receive a diagnosis of post-traumatic stress disorder (PTSD). Instead, his presentation would be coded as functional neurological disorder, from a chapter of the manual that has no formal relationship to trauma at all.

A partition with a date

Egyptian medical papyri from around 1900 BCE record tremors, paralysis, and aphonia without demonstrable lesion, attributed to displacement of the organs.^2,3 Through the long history of what would eventually be called hysteria, bodily dysfunction and disturbance of consciousness were treated as neighboring expressions of the same predicament. Jean-Martin Charcot’s formulation of traumatic hysteria in 1876 held them together explicitly: the dynamic lesion was not macroscopic injury but functional interruption produced by the representation of an event, and it manifested indifferently in paralysis, anesthesia, or amnesia.⁴ Pierre Janet’s synthesis of unintegrated memory and dissociation was built on the same continuity.^5,6

The partition arrives with the third edition of the Diagnostic and Statistical Manual in 1980.⁷ In establishing PTSD as a unified category organized around symptom clusters, DSM-III drew a boundary that placed conversion phenomena outside it. What had been one field became two: a trauma-related disorder defined by intrusion, avoidance, altered cognition, and hyperarousal, and a separate somatoform territory where bodily presentations were relocated and their traumatic genealogy quietly dropped.

This was not arbitrary, and it was not foolish. The 1980 codification was a genuine achievement. It removed from sufferers the charge of moral weakness that had followed them from the Munich Congress of 1916, where fear of insurance liability and of a weakened workforce weighed openly on a nosological decision,⁸ to the Second World War. It gave the condition a name that institutions and insurers could not dismiss. Precision of definition was part of what made that possible.

But precision has a cost, and the cost here was a class of patients.

What was lost, and where it went

The manual did not deny that conversion phenomena follow trauma, but simply stopped requiring anyone to ask about it. Somatoform and conversion diagnoses carry no etiological criterion; nothing in their application obliges a clinician to establish an exposure history, and nothing in their coding preserves the link if one is found. A patient whose paralysis follows torture and a patient whose paralysis follows nothing in particular receive the same label and, frequently, the same undirected care.

Something similar happened to a second group. The DSM-5 reintroduced a dissociative subtype for PTSD, restoring part of the older continuity, in line with converging clinical and neurobiological evidence that predated and prepared it.⁹ ICD-11 coding declined to follow, maintaining dissociative disorders as separate categories.¹⁰ The two systems now disagree about whether these phenomena belong inside the trauma field or beside it, which is a reasonable indication that the boundary is a convention rather than a discovery.

The disagreement extends further. Within the same population, lifetime prevalence estimated with DSM-5 criteria stands at around 11%, against a current-prevalence figure of 1.0% under ICD-11¹¹—figures that are not directly comparable, and whose contrast reflects the diagnostic architecture as much as the underlying condition.¹¹ The two manuals therefore do not simply generate different prevalence estimates; their differing diagnostic architectures can lead to different individuals being classified within, or outside, the PTSD spectrum. The divergence is best read not as an error on either side, but as two complementary operational strategies: one built for descriptive precision and the other for parsimony and rapid case definition.² Read either way, it tells us that the perimeter of the disorder is not being determined by nature alone.

Why an old boundary is a current problem

A historian could stop here. A clinician cannot, as the partition has begun to cost something it did not cost in 1980.

In many cultures the body is the principal register of distress.¹² Somatic presentations are not a failure to articulate psychological suffering; they are how suffering is articulated. European mental health services now receive, in growing numbers, patients from precisely those settings: people displaced by war, by persecution, by the collapse of the places they lived in. A classification that keeps bodily presentations outside the trauma core will systematically underrecognize them.

The mechanism is prosaic and therefore easy to miss. What is not asked about is not recorded. What is not recorded does not enter the case file. What is not in the case file does not appear in the aggregate data on which services are planned. A taxonomic decision taken for good reasons in one century becomes, in another, an instrument of exclusion.

The uses of an old question

The history of trauma is often read as a story of progressive clarification, running from moral condemnation through organic speculation to neurobiological understanding. Read that way, it flatters the present. Read more carefully, it shows something else: each era placed the wound where its own explanatory resources allowed—in the soul, in the womb, in the spinal cord, in the nerves, and in memory—and derived its treatments from each placement.²

This is not a relativist point. Some placements were better than others, and the current one has produced treatments that work. It is a point about self-awareness. Locating trauma in the cognitive-emotional sphere is the most recent in a long series of situated decisions, not the endpoint of the series. Historical materials do not prefigure our categories; they hold them up to a critical mirror.²

What that mirror shows, in this instance, is a specific question. Does the modern partition between the psychological and the somatic represent a genuine scientific advance, or has it obscured an older and more unitary understanding of embodied suffering? The question is not rhetorical, and it has empirical content; whether trauma exposure is systematically overrepresented among patients presenting with functional neurological symptoms is a question that can be asked of data, and one that current coding practice makes unnecessarily difficult to answer.

There is a modest remedy available while that work proceeds. It would not require dismantling the category or reopening the criteria. It would require only that trauma exposure be routinely asked about and recorded in patients presenting with functional neurological symptoms, and that where an association is found, it be preserved in the record rather than discarded at the moment of coding.

Epizelus is beyond our help. The patients who now arrive in outpatient clinics, unable to walk, see, or speak, who are asked about their bodies, but not about what happened to them, are not.

References

1. Herodotus. The Histories, book VI, 117.
2. Paganin W. Core symptoms of PTSD across four millennia: a phenomenological and nosographic analysis, from ancient Mesopotamian texts to modern psychiatric classifications. Med Humanit. 2026. doi:10.1136/medhum-2025-013623.
3. Okasha A. Mental health in Egypt. Isr J Psychiatry Relat Sci. 2005;42:116-125.
4. Micale MS. Charcot and les névroses traumatiques: scientific and historical reflections. J Hist Neurosci. 1995;4:101-119.
5. Janet P. Les médications psychologiques: études historiques, psychologiques et cliniques sur les méthodes de la psychothérapie. Paris: F. Alcan; 1919.
6. Craparo G, Cocco Ortu F, Van der Hart O, eds. Rediscovering Pierre Janet: Trauma, Dissociation, and a New Context for Psychoanalysis. New York: Routledge; 2019.
7. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. Washington, DC: American Psychiatric Association; 1980.
8. Holdorff B. The fight for “traumatic neurosis,” 1889-1916: Hermann Oppenheim and his opponents in Berlin. Hist Psychiatry. 2011;22:465-476.
9. Spiegel D, Loewenstein RJ, Lewis-Fernández R, et al. Dissociative disorders in DSM-5. Depress Anxiety. 2011;28:824-852.
10. Harrison JE, Weber S, Jakob R, et al. ICD-11: an international classification of diseases for the twenty-first century. BMC Med Inform Decis Mak. 2021;21:206.
11. Hoeboer CM, Nava F, Haagen JFG, et al. Epidemiology of DSM-5 PTSD and ICD-11 PTSD and complex PTSD in the Netherlands. J Anxiety Disord. 2025;110:102963.
12. Chentsova-Dutton Y, Maercker A. Cultural scripts of traumatic stress: outline, illustrations, and research opportunities. Front Psychol. 2019;10:2528.