Disease Classification

Summary

Disease classification — the systematic grouping of illnesses into categories — is one of medicine’s oldest projects and one of its most contested. From the Hippocratic Corpus through Enlightenment-era nosologies, nineteenth-century anatomical pathology, Kraepelin’s psychiatric taxonomy, and the modern DSM and ICD systems, physicians have repeatedly reorganized disease into new frameworks, each shaped by distinct assumptions about what a disease actually is. The central argument running through this history is that classification always encodes a theory: grouping diseases as natural species fixes them as things; grouping them as functional deviations treats them as states; grouping them by anatomical lesion locates them in space. Philosophers since Canguilhem have argued that pathology cannot be fully captured by any value-free system, because the judgment that a condition is a disease is always partly normative. Practitioners since Scudder have added a simpler complaint: even when classification works descriptively, it rarely tells you what to do therapeutically.

Ancient Classification

The Hippocratic Corpus was assembled in Alexandria in the third century BCE from texts written between 480 and 380 BCE by authors from several schools, containing mutually contradictory views.(Ackerknecht, 1955) The dominant Hippocratic approach was prognostic rather than classificatory: the traveling physician who could rapidly predict the course of an illness demonstrated the specialized knowledge that justified his fee and protected him from taking hopeless cases.(Ackerknecht, 1955)

The Methodist sect, which flourished from the first century BCE through the second century CE, proposed a radical simplification: all disease collapsed into three states — constriction, laxity, or a mixture of the two — requiring only three therapeutic responses. This reduction stripped out the elaborate Hippocratic attention to the patient’s constitution, season, and history, making the classification so streamlined that contemporaries mocked it. The Erasistratean tradition went in the opposite direction, developing a solidistic pathology that attributed disease to disturbances in solid parts rather than humors, anticipating later anatomical localism by nearly two millennia. These early debates established a permanent tension in disease classification between simplifying schemes that guide treatment and elaborating schemes that capture clinical reality.

Medieval Arab medicine made discriminations that persisted into modern classification.(Farr, William (Humphreys, Noel A., ed.), 1885) William Farr, reviewing the genealogy of his nineteenth-century nosology, credited the Arab physician Rhazes (died 923 CE) as the first to describe measles and scarlet fever under distinct names — a clinical separation that Avicenna, a century later, placed explicitly in his own classification by situating scarlet fever between smallpox and measles.(Farr, William (Humphreys, Noel A., ed.), 1885)

Sydenham and the Botanical Model

The medical classification of madness has its own distinct historical arc. Conrad and Schneider note that in ancient Hebrew culture, madness was attributed to divine retribution for sin; yet prophets who behaved in equally strange ways were not classified as mad but as divinely inspired — demonstrating that identical behaviors receive categorically different social labels depending on perceived social function.(Peter Conrad and Joseph W. Schneider, 1980) Greek medicine rejected the supernatural framework and produced the first naturalistic disease classification for mental states through humoral theory: Hippocrates attributed the different forms of madness to imbalances of the four humors, and the terms he coined — epilepsy, mania, melancholia, paranoia — remain in use today, marking the extraordinary persistence of categories invented within one theoretical framework long after that framework has been abandoned.(Peter Conrad and Joseph W. Schneider, 1980)

Thomas Sydenham came to medicine more by accident than design, fully sharing in the military and political vagaries of his family’s fortune.(Dewhurst, 1966) His Puritanism caused him to rebel against all that was useless in orthodox medicine.(Dewhurst, 1966) Dewhurst characterizes Sydenham’s great merit as avoiding both the iatrophysical school (body as machine, following Galileo and Descartes) and the iatrochemical school (body as chemical reactions, following Paracelsus and van Helmont) as well as Galenic orthodoxy, instead concentrating on perfecting clinical practice through a “plain, historical approach to clinical problems.”(Dewhurst, 1966)

From that observation Sydenham developed his botanical model of disease: just as the Supreme Being produced animals and plants according to fixed species, so diseases followed structural laws that held across individuals. Each disease was a natural kind with an observable course — characteristic onset, characteristic sequence of symptoms, characteristic resolution — that the physician could read once he learned to look past individual patient variation. Foucault describes this precisely: in classificatory medicine, “disease is given an organization, hierarchized into families, genera, and species” independent of any localization in a particular body.(Foucault, 1963) The body was secondary; the disease-species was primary. To know the truth of a pathological condition, the physician had to “abstract the patient” from his individual constitution and focus on what was essential to the disease itself.(Foucault, 1963)

Yet Sydenham’s concept of the “epidemic constitution” pulled against the pure botanical model. Each year’s disease patterns were shaped by something in the atmospheric and telluric environment — not a miasma exactly, but a shifting complex of natural conditions that modified how individual diseases manifested.(Foucault, 1963) This meant disease species were not entirely stable across time and place; the same nosological category might behave differently under different constitutions.(Foucault, 1963) Quinine’s arrival from Peru in the 1630s applied pressure from a different direction: the drug cured malaria without producing any of the evacuations Galenist pathology claimed necessary, undermining the humoral framework.(Ackerknecht, 1955)

William Farr, writing in the mid-nineteenth century, traced the zymotic hypothesis of epidemic disease — the idea that infectious diseases were driven by specific ferment-like agents — back to Sydenham, Thomas Morton, and Thomas Willis as three contemporaries who announced it in the seventeenth century, well before bacteriology gave it a material basis.(Farr, William (Humphreys, Noel A., ed.), 1885) Sydenham’s role in this lineage was thus double: his botanical model imposed specificity on disease categories, and his epidemic-constitution thinking pointed toward the environmental triggers that would later be conceptualized as disease agents.

Enlightenment Nosologists

Sydenham’s botanical program found its most systematic elaboration in François Boissier de Sauvages’s Nosologia Methodica (1761), which organized 2,400 disease species into ten classes, 295 genera, and thousands of species on the model of Linnaean botany. William Cullen, working in Edinburgh, took a different approach: rather than multiply species, he sought to reduce all pathology to a single mechanism. Cullen held that the nervous system was the key to the “animal economy,” that all pathology originated in disordered nervous action — “spasm” — and he organized disease into four principal classes on that basis.(Porter, 1997) Both Sauvages and Cullen represent what Foucault calls the “primary spatialization” of disease: defining disease in an abstract classificatory space of families and genera, prior to any localization in a particular body.(Foucault, 1963)

In classificatory medicine, disease was perceived in a space of projection without depth, of coincidence without development, making its epistemological logic spatial and synchronic.(Foucault, 1963) The patient was treated as a disturbance that had to be subtracted to reveal the disease’s pure essence; the disease had an identity independent of the patient, who was its secondary support rather than primary reality.(Foucault, 1963) Classificatory medicine used a botanical model in which diseases were conceived as natural species with fixed essences, following the same structural laws as plants and animals.(Foucault, 1963)

The new style of totalization replaced the closed treatises of the eighteenth century with open, infinitely extendable tables of medical events.(Foucault, 1963) Instead of mapping symptoms onto a nosological plan, medical knowledge now carved up a field through the interplay of intersecting series that reconstitute a chain.(Foucault, 1963) This shift was not merely methodological; it registered a new understanding of disease as an event in a chain of intersecting causes rather than a species instantiated in a body.(Foucault, 1963)

Farr’s own Statistical Nosology — presented in the second annual report of the Registrar-General (1837) and revised across several decades — traced its lineage from Sauvages through Linnaeus, Vogel, Sagar, Selle, Ploquet, Crichton, Macbride, Darwin, and Cullen, continuing to Pinel, Richerand, Bichat, and Mason Good.(Farr, William (Humphreys, Noel A., ed.), 1885) This historiography was not mere erudition; it positioned statistical nosology as the culmination of over a century of systematic classification, now given a new purpose: converting disease categories into countable and comparable units.

Paris School and Anatomical Localism

The Paris hospital medicine that emerged from the Revolutionary reorganization introduced new programmes of medical inquiry, new disease concepts, and new research practices, culminating in a distinctive paradigm characterized by scientific observation raised on pathological anatomy, the paradigm of the lesion, quantification, and clinical-pathological correlation.(Porter, 1997)

Giovanni Battista Morgagni’s On the Sites and Causes of Disease (1761), published when he was seventy-nine, shifted emphasis from humors to localized organ changes and causally connected those changes with clinical symptoms, reconceiving disease as a lesion in a specific anatomical site.(Ackerknecht, 1955) Paris hospital medicine, characterized by scientific observation raised on pathological anatomy, the paradigm of the lesion, quantification, and clinical-pathological correlation, further advanced this approach.(Porter, 1997) Pierre Louis’s numerical method introduced arithmetic to test therapies against outcomes, revealing, for example, that bloodletting made no measurable difference to pneumonia outcomes regardless of timing or volume.(Porter, 1997)

The localistic paradigm had an important consequence for surgery. As long as humoral theory held, removing a tumor was irrational — the tumor expressed a systemic dyscrasia and would regrow at the same or another site. Once disease was understood as residing in a local lesion, surgical removal became conceptually coherent, and surgical activity increased significantly in the decades before anesthesia and antisepsis.(Ackerknecht, 1955)

Charles Rosenberg, in his 2002 analysis of diagnostic culture, argues that this shift was largely accomplished by the 1860s and that germ theory — so often credited with transforming disease-entity thinking — was not the driver but the beneficiary of a conceptual transformation already complete.(Rosenberg, Charles E., 2002) Pathological anatomy, physical diagnosis, and chemical pathology together produced stable disease entities conceived as independent of any individual body; germ theory then gave those entities a causal story. Richard Bright’s disease, named in the 1820s and defined through chemical pathology (albumen in heated urine) and post-mortem kidney findings, was the archetype: a “doctor’s disease,” configured not around the patient’s felt experience but around laboratory findings and autopsy correlation.(Rosenberg, Charles E., 2002)

German clinicians in the 1840s resisted the localistic paradigm from a physiological direction. Their slogan was “pathological physiology”: disease was not a thing located in an organ but a disturbance of function — a process, not a site.(Ackerknecht, 1955) Claude Bernard, elaborating the physiological approach, declared the laboratory the “sanctuary” of medicine, situating experimental physiology as a distinct era in medicine’s development — the successor to the library, the bedside, and the hospital.(Ackerknecht, 1955) They rejected what they called the “ontological approach” of the Paris and Vienna schools. The irony Ackerknecht notes is that Carl Wunderlich, who set out explicitly to fight ontological disease-entity thinking through his systematic temperature charts, ended by constructing an ontology of his own: “specific temperature charts for specific diseases” were as entity-like as anything he sought to oppose.(Ackerknecht, 1955) The tension between classifying diseases as distinct entities and understanding them as functional states on a continuum was not resolved by the physiological critique; it was relocated.

Broussais had already pushed against localism from within the Paris school, arguing that disease was not ontologically distinct from health but occurred when normal functions went awry — “shades of grey” rather than black and white.(Porter, 1997) Auguste Comte elevated Broussais’s nosological principle to a universal axiom applicable to biological, psychological, and sociological phenomena.(Canguilhem, 1966) In Comte’s formulation, the pathological state differed from the physiological only in intensity, never producing truly new phenomena.(Canguilhem, 1966)

William Farr and Statistical Nosology

William Farr (1807-1883), compiler of abstracts in the General Register Office from 1839 onwards, built the most consequential nosological programme of the nineteenth century — not a philosophical taxonomy but a working instrument for measuring population mortality and tracking disease across time and space. His core methodological claim was that nomenclature was as essential to vital statistics as standard weights and measures were to the physical sciences.(Farr, William (Humphreys, Noel A., ed.), 1885) He had stated this argument as early as his First Annual Report (1839), where he also catalogued how every disease had been denoted by multiple confusing and overlapping terms requiring systematic standardization before any comparative statistics could be drawn.(Farr, William (Humphreys, Noel A., ed.), 1885) Without stable, agreed disease categories, mortality data from different districts, different years, and different countries could not be compared; statistics would remain an inventory of incompatible local descriptions rather than a science of population health. Rosenberg, citing Farr’s 1837 formulation directly, identifies this argument as the moment when epidemiology became structurally dependent on standardised disease categories — and thus committed to treating diseases as things that exist independently of the individual patients who carry them.(Rosenberg, Charles E., 2002)

Farr organized prior nosological traditions into three methodological types — anatomical, physiological, and topographical — and constructed his own Statistical Nosology around five primary divisions.(Farr, William (Humphreys, Noel A., ed.), 1885) In situating this work within medicine’s long history, he defined inflammation following the classical Celsus criteria of heat (calor), redness (rubor), swelling (tumor), and pain (dolor), anchoring his nosological project in the ancient clinical tradition even while advancing modern statistical classification.(Farr, William (Humphreys, Noel A., ed.), 1885) Class I (Zymotici) covered epidemic, endemic, and contagious diseases; Class II (Cachectici) covered constitutional diseases including gout, dropsy, cancer, and scrofula; Class III (Monorganici) covered local diseases of the nervous, circulatory, respiratory, digestive, urinary, reproductive, locomotive, and integumentary systems; Class IV (Metamorphici) covered developmental diseases from premature birth through old age; and Class V (Thanatici) covered violent deaths.(Farr, William (Humphreys, Noel A., ed.), 1885) Within Class I, Farr further distinguished four orders: Miasmatic diseases; Enthetic diseases (transmitted by contact or inoculation); Dietetic diseases (scurvy, ergotism); and Parasitic diseases.(Farr, William (Humphreys, Noel A., ed.), 1885)

The term “zymotic” — from the Greek zymoo, to ferment — was Farr’s own coinage, proposed to the International Statistical Congress for adoption as a standard nosological term.(Farr, William (Humphreys, Noel A., ed.), 1885) It carried a theoretical commitment: zymotic diseases were driven by specific agents analogous to chemical ferments, whose composition was unknown but whose existence was demonstrated by their effects, following Liebig’s fermentation analogy.(Farr, William (Humphreys, Noel A., ed.), 1885) The nomenclature was systematic to the point of proposing names for the individual disease agents — varioline for smallpox, typhine for typhus, cholerine for cholera, syphiline for syphilis — an entire lexicon of disease exciters constructed before bacteriology had identified any of them.(Farr, William (Humphreys, Noel A., ed.), 1885) Farr applied the same naming logic to specific disease episodes: when discussing cholera transmission in his statistical reports, he proposed the term “cholrine” for the cholera-causing matter, maintaining the systematic practice of naming specific agents even when their nature was disputed.(Farr, William (Humphreys, Noel A., ed.), 1885) The project illustrates how nosological frameworks can run ahead of etiological knowledge, creating named entities that await empirical filling-in.

Farr reported that the Royal College of Physicians published disease nomenclature in Latin, French, Italian, and German, underscoring the international ambition of standardized disease classification.(Farr, William (Humphreys, Noel A., ed.), 1885) The Compulsory Medical Certification Act 1874 made medical certification of death mandatory in England, a legislative milestone that Farr had advocated for decades.(Farr, William (Humphreys, Noel A., ed.), 1885) Furthermore, diphtheria was separated from scarlet fever in the English registration returns in 1855, and enteric fever was distinguished from typhus in 1869, illustrating how the disease classification system evolved incrementally.(Farr, William (Humphreys, Noel A., ed.), 1885)

Farr also cited the historian Niebuhr to justify the classification’s ambition: epidemic diseases “distinguished one country from another, formed epochs in chronology,” which meant that systematic statistical surveillance of epidemic patterns was not merely a medical undertaking but a form of historical knowledge about national populations.(Farr, William (Humphreys, Noel A., ed.), 1885)

The registration system that Farr built with these categories also documented its own incompleteness. In one year’s English data, 4,630 deaths had no cause specified and 3,506 were inferred sudden deaths; the Royal College of Physicians responded by publishing its disease nomenclature in Latin, French, Italian, and German, asserting the international ambition of standardized terminology even as local registration quality remained uneven.(Farr, William (Humphreys, Noel A., ed.), 1885) The Compulsory Medical Certification Act of 1874 eventually made cause-of-death certification mandatory in England — a legislative landmark that Farr had advocated for decades.(Farr, William (Humphreys, Noel A., ed.), 1885) The incremental refinement of the classification system itself appears in the registration record: diphtheria was separated from scarlet fever in the English returns in 1855, and enteric fever was distinguished from typhus in 1869.(Farr, William (Humphreys, Noel A., ed.), 1885) Each such separation required both a clinical argument that the conditions were distinct entities and an administrative decision to introduce a new category into the recording apparatus. The registrations also enabled Farr to document that different infectious diseases had distinctly different relationships to age-related susceptibility: his analysis of disease age-distributions showed a mean age at death of 1.8 years for whooping cough, 2.7 years for measles, and 5.8 years for scarlatina and diphtheria — an empirical pattern that could not have been discerned without the standardized classification system he had established.(Farr, William (Humphreys, Noel A., ed.), 1885)

Farr’s commitment to systematic nomenclature extended beyond disease to personal names. Analysing two quarterly registration indexes, he found 32,818 distinct surnames and estimated that England and Wales had between 35,000 and 40,000 distinct surnames in total — the first large-scale empirical study of English family nomenclature, produced as a by-product of the same registration machinery that generated the disease statistics.(Farr, William (Humphreys, Noel A., ed.), 1885) This was not mere curiosity: accurate personal identification mattered to the registration system’s ability to link individual records across time and space.

The clinical separation of “vesanic” (functional, psychiatric) delirium from “non-vesanic” (organic, fever-related) delirium created the conceptual foundation for the modern notion of psychosis as a distinct category.(German E. Berrios & Roy Porter (eds.), 1995) Vesanic delirium referred to those states occurring in the course of mental illness proper, while non-vesanic delirium was reserved for those associated with physical disease.(German E. Berrios & Roy Porter (eds.), 1995)

Psychiatric Classification

Psychiatric nosology developed along a partly independent track, shaped by the particular institutional context of asylum medicine and the particular problem of whether mental illness was a real medical entity at all. The conceptual precondition for any psychiatric nosology was the Enlightenment shift that recognized mental disorder as a disease rather than a form of possession, sin, or crime; this recognition enabled a more humane approach to institutional care, most visibly in Philippe Pinel’s removal of chains from his insane patients at the Bicetre hospital in Paris in 1794.(Ackerknecht, 1955) Antoine-Laurent Bayle’s 1822 doctoral thesis, which attributed the psychiatric symptoms of neurosyphilis to chronic meningeal inflammation, was the first demonstration of a psychiatric disease with definite organic substrate — establishing a model in which psychiatric conditions would eventually be decoded as brain lesions.(Shorter, 1997)

The model proved harder to generalize than expected. Wilhelm Griesinger in Berlin declared in the 1860s that “patients with so-called mental illnesses are really individuals with illnesses of the nerves and brain,” but the brain lesions that would confirm this declaration remained elusive for most conditions.(Shorter, 1997) The effort to classify psychiatric illness anatomically — mapping specific symptoms to specific brain regions — produced, in Karl Jaspers’s phrase, “brain mythology” rather than clinical science.

Delirium played an important conceptual role in this history. Rather than being classified alongside other forms of physical illness, delirium served as the primary clinical model from which the modern notion of psychosis was constructed.(German E. Berrios & Roy Porter (eds.), 1995) The clinical separation of “vesanic” (functional, psychiatric) delirium from “non-vesanic” (organic, fever-related) delirium created the conceptual foundation for psychosis as a distinct category independent of observable brain pathology.(German E. Berrios & Roy Porter (eds.), 1995) The transformation of delirium from a fever-associated behavioral syndrome into a disorder of consciousness, attention, and orientation — shaped by changing concepts of consciousness itself — was a process internal to psychiatric classification rather than imposed from outside by pathological anatomy.

Emil Kraepelin’s longitudinal, prognosis-oriented approach marked the decisive shift. Where his contemporaries classified psychiatric patients by cross-sectional symptom pattern — what the patient had at a given moment — Kraepelin tracked illness over time, finding that the course of illness offered a sharper guide to its nature than any symptom profile observable at a single encounter.(Shorter, 1997) This longitudinal method produced his fundamental two-category system: dementia praecox (subsequently schizophrenia) as a deteriorating condition, and manic-depressive insanity as a non-deteriorating one. The SEP entry on disease concepts identifies this as an application of the “symptom-cluster model” to psychiatry — a direct descendant of Sydenham’s botanical program, now organized by prognosis rather than observable symptom pattern.(Dominic Murphy, unknown)

The political dimension of psychiatric classification became unavoidable in the twentieth century. Samuel Cartwright proposed “drapetomania” in 1843 as a disease causing enslaved people to run away; Soviet psychiatrists diagnosed political dissidents with “sluggish schizophrenia”; the DSM listed homosexuality as a disorder until 1973.(Dominic Murphy, unknown) These cases are central to constructivist critiques of psychiatric classification but are resisted by naturalists as examples of incorrect application of an otherwise valid concept — a debate that continues in the current DSM controversy.

Conrad and Schneider situate the DSM within a longer argument: disease categories are not discovered but constructed through political processes, with the deviance designation (sin, crime, or sickness) determining which institution — church, state, or medicine — holds legitimate social control authority.(Peter Conrad and Joseph W. Schneider, 1980) Illness, on their account, is “a social construction based on human judgments of some condition in the world” — not a natural kind but an evaluative act that always carries a negative moral value, since no illness designation has ever been a positive judgment.(Peter Conrad and Joseph W. Schneider, 1980) Biological aberration is neither necessary nor sufficient for illness designation: a seven-foot basketball player is outside the normal biological range but not considered ill, while most functional mental disorders have no or at best questionable physiological evidence yet are defined and treated as diseases.(Peter Conrad and Joseph W. Schneider, 1980) DSM-I (1952) classified homosexuality under “Sociopathic Personality Disturbance” and DSM-II (1968) gave it specific designation 302.0 under “Sexual Deviation”; in December 1973 the APA Board of Trustees voted to remove homosexuality per se from the nomenclature, and in the subsequent membership referendum 58% of voting members endorsed the change — a decision made by vote, not by scientific discovery, illustrating that psychiatric classification responds to political and social pressures as much as to evidence.(Peter Conrad and Joseph W. Schneider, 1980)(Peter Conrad and Joseph W. Schneider, 1980) The medical model of deviance is “assumed to have a scientific basis and thus treated as if morally neutral” — but medical language “obfuscates” the moral judgments embedded in disease designations: to call something a disease is to deem it undesirable, whatever the technical vocabulary suggests.(Peter Conrad and Joseph W. Schneider, 1980)(Peter Conrad and Joseph W. Schneider, 1980) Conrad and Schneider trace the longer historical arc: the transformation of deviance from “badness” (sin or crime) to “sickness” is not a natural evolution but a profoundly political process, representing the ascending pattern for deviance designations in post-industrial Western societies.(Peter Conrad and Joseph W. Schneider, 1980)

The history of 19th-century disease classification reveals how medical categories encode moral judgments about sexuality. 18th-century theories of the body as a closed system of vital nervous energy provided the first medical grounding for pathologizing deviant sexuality: activities that made “repeated, unusual, and ‘unhealthy’ (immoral) demands on one’s body would lead inevitably to its depletion, debility, wasting, and disease” — translating moral disapproval of sexual behavior into a claim about physiological mechanism, and making morally transgressive conduct a medical category without changing its normative content.(Peter Conrad and Joseph W. Schneider, 1980) The anti-masturbation discourse of the 18th and 19th centuries functioned as a covert classification system for same-sex conduct: the terms “onanism” and “the secret sin” were used to encompass a variety of deviant sexual practices, making the chronic masturbator’s clinical portrait — effeminate, depleted, socially disabled — a direct prototype for the later psychiatric classification of the homosexual.(Peter Conrad and Joseph W. Schneider, 1980) Similarly, the 1898 Bayer Laboratory introduction of heroin as a “non-addicting substitute” for morphine illustrates how pharmaceutical classification determines moral status: once heroin’s addictive properties became clear and physicians withdrew medical legitimacy from it, the drug shed its clinical category and acquired exclusively recreational status, becoming “quickly imbued with far greater negative connotations than either of its sister drugs” despite equivalent pharmacology.(Peter Conrad and Joseph W. Schneider, 1980)

Biological criminology extended the classification project to behavior itself. Cesare Lombroso’s 1876 theory of the “atavistic being” — the criminal as a biological throwback identifiable by physical stigmata — represented the application of taxonomic thinking to human deviance: just as botanical nosology organized plants into natural kinds, Lombroso sought to organize criminals into biological types identifiable by measurement.(Peter Conrad and Joseph W. Schneider, 1980) The XYY chromosome theory of the 1960s applied the same logic with modern genetics: Patricia Jacobs’s 1965 finding of 3.5% XYY among Scottish prison inmates against 0.13% in the general population appeared to provide a chromosomal basis for a criminal classification, though the claim did not survive replication.(Peter Conrad and Joseph W. Schneider, 1980) Alfred Lindesmith’s sociological counter-model showed why purely biological classification fails for complex behavioral states: addiction occurs, he argued, only when opiates are used to alleviate withdrawal distress “after such distress has been properly understood or interpreted” — if the individual does not cognitively recognize their distress as withdrawal, they do not become addicted regardless of the drug’s pharmacological action.(Peter Conrad and Joseph W. Schneider, 1980) Lindesmith, Becker, and Young more broadly demonstrated that “repeated drug taking is not the automatic consequence of the drug’s pharmacological properties but rather a complex process wherein the individual learns to use the drug under particular circumstances” — making addiction a social learning phenomenon that classification systems organized around biological mechanisms cannot capture.(Peter Conrad and Joseph W. Schneider, 1980)

Conrad and Schneider also challenge the narrative that 19th-century public health improvements reflected clinical medical progress. The dramatic reduction in epidemic disease was “by and large not the result of new medical knowledge or improved clinical medical practice” but of rising living standards, better nutrition and housing, and sanitation innovations: “With the lone exception of vaccination for smallpox, the decline of these diseases had nearly nothing to do with clinical medicine” — yet medicine received popular credit, providing cultural authority for the expansion of medical classification into social life.(Peter Conrad and Joseph W. Schneider, 1980)

Statistical and Actuarial Classification

Adolphe Quetelet’s development of the “average man” (l’homme moyen) in the 1830s introduced a new basis for disease classification: the statistical population norm.(Porter, 1997) Where Broussais and Bernard had argued that disease was an intensified version of normal physiology, Quetelet provided the conceptual infrastructure for defining the normal by its statistical distribution — and the abnormal as deviation from that distribution. This shift, as the SEP entry on disease concepts describes it, moved the concept of disease from the Sydenham/Kraepelin symptom-suite model and the localist pathological-process model toward an “actuarial” model: elevated statistical risk itself counts as disease even in the absence of overt symptoms.(Dominic Murphy, unknown) Hypertension, impaired glucose tolerance, and elevated cholesterol are disease states in this model not because they produce symptoms but because they elevate the probability of future events.

Rosenberg identifies the actuarial model’s full emergence as a product of technology-enabled screening, which created what he calls “proto-diseases” — conditions such as elevated cholesterol, low bone density, or carrier status for genetic diseases that occupy a position between warning signal and pathology.(Rosenberg, Charles E., 2002) These occupy a structurally novel position: they alter clinical and social life, trigger treatment decisions, and generate diagnostic identities before any symptom has appeared. The proto-disease category exposes a feedback loop between classification and clinical practice: once a screening technology exists to detect a quantitative deviation, the category of disease expands to include it, and classification drives the generation of new patient populations rather than merely organizing the populations that present spontaneously.

The International Classification of Diseases (ICD), first assembled by the International Statistical Institute in the 1890s, institutionalized statistical thinking about disease categories on a global scale. Its iterations have been as much political negotiations as scientific classifications: decisions about which conditions to include, how finely to subdivide, and what causal principles to recognize reflect the interests of pharmaceutical companies, disability advocates, insurance systems, and professional organizations alongside the clinical observations of physicians.

The Tyranny of Diagnosis: Rosenberg’s Analysis

Charles Rosenberg’s 2002 article “The Tyranny of Diagnosis: Specific Entities and Individual Experience,” published in The Milbank Quarterly, provides the most systematic recent analysis of disease specificity as a historical and social formation. His central claim is that the modern history of diagnosis is “inextricably related to disease specificity” — to the idea that diseases are entities that exist outside any particular patient’s experience of illness.(Rosenberg, Charles E., 2002) In traditional (pre-nineteenth-century) medicine, he argues, disease concepts were organised around the individual sufferer: they were “symptom oriented, fluid, idiosyncratic, labile, and prognosis oriented,” focused on transient moments in a bodily process rather than on stable categorical entities.(Rosenberg, Charles E., 2002) The shift to mechanism-based specificity was largely in place by the 1860s — before germ theory — driven by pathological anatomy, physical diagnosis, and chemical pathology.(Rosenberg, Charles E., 2002)

Rosenberg argues that disease entities acquire social reality through three interlocking factors: technology (diagnostic and therapeutic innovations that define and alter diseases), the hospital as a research and care site, and bureaucratic structures including insurance coding and nosological tables.(Rosenberg, Charles E., 2002) The acute-care hospital’s internal administrative order required diagnostic categories; case-record forms that demanded a diagnosis while leaving little space for patient narrative embodied this orientation.(Rosenberg, Charles E., 2002) The hospital was not merely a site where disease was treated but a machine for producing and stabilizing disease entities as administrative realities.

The pedagogical apparatus of clinical medicine contributed to the same process. Rosenberg traces the term “differential diagnosis” — now central to clinical teaching — to Richard Cabot’s early twentieth-century didactic efforts, noting that the adjective “differential” presupposes discrete alternatives to differentiate among, thereby legitimating and prospectively creating the entities among which the clinician is choosing.(Rosenberg, Charles E., 2002) The concept is circular in a productive sense: teaching physicians to think in terms of discrete differential possibilities both describes and reinforces a world organized by disease entities. Similarly, pernicious anemia was definitionally constituted as a specific disease in the 1920s by its response to liver extracts; the predictability of therapeutic response implied a specific pathological mechanism and thereby conferred epistemological legitimacy on the category.(Rosenberg, Charles E., 2002) Therapeutic response and disease identity became mutually constitutive.

Diagnosis also functions, Rosenberg argues, as a ritual that links the individual patient to the collective medical system — legitimating medical authority, facilitating clinical decisions, and providing culturally agreed-upon meanings for suffering.(Rosenberg, Charles E., 2002) When diagnosis operates in this register, it is indispensable: it orders anxiety, names what is happening, and connects the sick person to institutions capable of providing care. In the act of diagnosis, however, the patient is necessarily objectified, “recreated into a structure of linked pathological concepts and institutionalized social power” — a bureaucratically alienated disease-defined self that exists in the institutional software of protocols and reimbursement codes.(Rosenberg, Charles E., 2002) Rosenberg’s title captures both sides: “tyranny” and “indispensability” are not contradictory characterizations but descriptions of the same phenomenon from different angles.(Rosenberg, Charles E., 2002)

A further line of Rosenberg’s analysis traces the recurring use of specific-mechanism framing to manage social deviance. Neurasthenia, hysteria, sexual psychopathy, homosexuality, and alcoholism have each been framed in the language of disease specificity — claiming value-neutral mechanistic status while serving normative cultural purposes.(Rosenberg, Charles E., 2002) The specific-entity framework consistently performs this double function: it claims to describe the natural world while organizing the social world. This analysis extends the constructivist critique of psychiatric diagnosis to all of medicine, locating the political dimension not in psychiatric nosology specifically but in the logic of specificity as such.

Against what he calls the “tyranny” of diagnosis, Rosenberg holds open an alternative: illness and disease are not closed opposing systems but “mutually constitutive and continuously interacting worlds.”(Rosenberg, Charles E., 2002) The patient’s lived experience of suffering cannot be fully captured by medical categories — but it is never entirely separate from them either. Classification is indispensable; its tyranny lies in the tendency to treat its categories as more stable and complete than they are.

Philosophical Critiques

The most sustained philosophical critique of disease classification came from Georges Canguilhem, whose 1943 doctoral thesis in medicine — published in English as The Normal and the Pathological — argued that the entire nineteenth-century quantitative program, from Broussais and Comte through Claude Bernard, was built on a confusion.(Canguilhem, 1966) Canguilhem argued that the pathological state does not consist in the absence of norms but is itself a norm of life — an inferior norm that tolerates no deviation from the conditions in which it is valid.(Canguilhem, 1966) The sick person is not abnormal because of the absence of a norm but because of incapacity to be normative — to establish other norms under new conditions. This means the pathological cannot be defined as a quantitative deviation from the physiological, because it is a qualitatively different mode of life.

Canguilhem also challenged the coherence of “objective pathology” as a category: one can carry out objectively impartial research in pathology, he argued, but pathology’s object cannot be conceived without being related to positive and negative qualifications — it is not a fact but a value.(Canguilhem, 1966) There is, strictly speaking, no biological science of the normal as such; there is only a science of biological situations and conditions called normal, which is physiology.(Canguilhem, 1966)

Foucault extends this argument historically. The exact superposition of the body of the disease and the body of the sick person — the assumption that disease is located in the individual body — is a historically contingent formation specific to nineteenth-century anatomical medicine.(Foucault, 1963) The Paris school’s paradigm of the lesion was not a discovery of what disease always was; it was an epistemological reorganization that installed a new relationship between medicine’s objects and its techniques of observation.

The eclectic physician John Milton Scudder anticipated both critiques in practical terms. Writing in 1883, he observed that the Royal College of Physicians classified 1,146 distinct diseases, while physicians employed only about one hundred remedies, with five to ten in common use for everything.(Scudder, 1883) Classification, he argued, was useful for studying pathology but harmful when made the basis of therapeutics: it carried the false idea of disease as an entity — a thing that had precise form and condition — rather than a condition of life.(Scudder, 1883) The physician who thought in disease categories would try to expel the disease rather than restore function; he would forget the patient’s life in his effort to rid him of his disease.(Scudder, 1883) Scudder proposed instead a four-part classification: by causation (epidemic, contagious, endemic, sporadic), by scope (general or local), by tissue change (structural or functional), and by physiological deviation (excess, defect, or perversion of function).(Scudder, 1883) The purpose was not a comprehensive nosology but a minimal framework adequate to guide remedy selection.

Contemporary philosophy of medicine maps the debate between naturalism (disease as objective biological malfunction) and constructivism (disease as socially valued judgment about biological conditions) across a 2x2 matrix of positions rather than a simple binary.(Dominic Murphy, unknown) Christopher Boorse’s biostatistical theory defines disease as departure from species-typical design, distinguishing “disease” (biological malfunction) from “illness” (a disease judged undesirable). Jerome Wakefield’s “harmful dysfunction” analysis adds an evolutionary criterion: a condition is a mental disorder only if a psychological mechanism fails to perform its naturally selected function and this failure is harmful to the person. Stegenga’s taxonomy identifies four positions: naturalism, normativism, hybrid accounts, and eliminativism — the last arguing for replacing the disease concept entirely with successor notions tied more closely to the relevant sciences.(Dominic Murphy, unknown)

See Also

• nosology
• pathological-anatomy
• humoral-theory
• biological-normativity
• paris-clinical-school
• thomas-sydenham
• emil-kraepelin
• georges-canguilhem
• dsm
• vitalism
• disease-as-entity
• epidemic-constitution
• william-farr
• vital-statistics
• zymotic-disease

Sources

Editorial Notes

Gaps the encyclopaedia compiler flagged for future evidence work, collected from inline markers in the body and frontmatter.

Statistical and Actuarial Classification

• [GAP: specialist source needed — ICD political and negotiation history requires Bowker and Star’s Sorting Things Out (1999) or WHO archival scholarship not in Library]

Conrad-Schneider additions (2026-05-04)

• Conrad, P., & Schneider, J. W. (1980). Deviance and Medicalization: From Badness to Sickness. St. Louis: Mosby. [Source ID: conrad-schneider-deviancemedicalization-1980]