Depression is now the most common psychiatric diagnosis in the world, applied to hundreds of millions of people across every culture. But the word itself is less than two centuries old, the disease category it names was assembled only in 1980, and the condition it replaced — melancholia — had an entirely different history. Understanding depression means understanding how a word borrowed from cardiovascular medicine, a tradition reaching back to ancient Greece, a dispute among twentieth-century nosologists, and the commercial interests of pharmaceutical companies all converged on the same diagnostic label.
The Ancient Substrate: Melancholia
The conceptual history of depression begins not with the word but with its predecessor. The classical Greek term melancholia denoted a chronic mental illness whose defining features were causeless sadness and fear, arising from excess of black bile (melaina chole), the cold and dry humor that dominated the autumn season.(Lawlor, 2012) Black bile, in the humoral system inherited from the Hippocratics, was one of four bodily fluids — alongside blood, yellow bile, and phlegm — that mapped onto the Pythagorean elements, the seasons, and the cardinal qualities of hot, cold, wet, and dry.(Lawlor, 2012) Under this somatic framework, black bile (cold and dry) produced melancholy, while excess blood or yellow bile produced mania — somatic explanations that governed Western understanding of mood disorders for roughly two millennia, from Hippocrates through the Renaissance.(German E. Berrios & Roy Porter (eds.), 1995) Melancholia in this framework was not a specific disease entity but the result of humoral imbalance; its severity varied with the degree of imbalance, and the boundary between ordinary constitutional variation and frank illness was one of degree rather than kind.(Lawlor, 2012)
Galen elaborated this framework by identifying three distinct subtypes of melancholy based on where the excess black bile accumulated: in the brain directly, generalized in the blood (darkening the skin), and in the hypochondrium (the region just below the ribcage), where digestive disturbance produced atrabilious vapors that ascended to the brain.(Lawlor, 2012) The hypochondriacal type — producing flatulence, digestive complaints, and depression together — would generate a family of diagnoses that persisted through the seventeenth and eighteenth centuries. Galen’s explanation of why black bile induced fear was itself a physical metaphor: the darkness of the humor cast a shadow over the “area of thought” in the brain, just as external darkness renders people fearful.(Lawlor, 2012)
Melancholia was not, before the nineteenth century, primarily a disorder of sadness. Berrios’s careful historical analysis establishes that up to the period of the Napoleonic Wars, melancholia was a broad category encompassing states whose only common denominator was the presence of few delusions — as opposed to many, which would indicate mania. Sadness and low affect were not considered defining criteria. Cases that today would receive a diagnosis of schizophrenia were routinely classified as melancholia.(German E. Berrios & Roy Porter (eds.), 1995) This is one of the most important correctives to retrospective readings of the history: what the Hippocratics and Galenists called melancholia was not what we call depression.
The Pseudo-Aristotelian Problemata XXX (c. 350 BC) introduced a countervailing tradition that would run for two millennia: the melancholic as genius. The text asked why all outstanding men in philosophy, poetry, politics, and the arts are melancholic.(Lawlor, 2012) This identification of black bile excess with intellectual and creative gifts established a framework that the Renaissance would elaborate into an explicit ideology of melancholic genius, shaping everything from Dürer’s Melencolia I to Shakespeare’s Hamlet. Cultural representations consistently split the maniac and the melancholic along a class and value axis: where the maniac occupied the role of the subhuman “other,” dangerous and degraded, the melancholic was the gifted genius — a polarity traceable from Aristotle’s Problem XXX through Renaissance Neoplatonism to Romantic Weltschmerz.(German E. Berrios & Roy Porter (eds.), 1995)
Islamic Medicine: Al-Balkhi’s Classification
Before the Renaissance, and running in parallel with the Galenic tradition’s transmission through Islamic medicine, the ninth-century Central Asian physician Abu Zayd al-Balkhi developed what may be the earliest systematic classification of depressive states. Working in Baghdad, al-Balkhi distinguished three categories: normal sadness (huzn) that afflicts all people in response to life’s losses; reactive depression with a known cause (loss of a loved person, bankruptcy, bereavement); and endogenous depression with no known reason, which he attributed to bodily factors such as impurity of the blood.(Malik Badri, 2013) In his analysis, al-Balkhi defined sadness as an emotion directed at the past and fear as one directed toward the future, identifying them as the strongest among psychological symptoms: when they afflict a person simultaneously, they produce the worst form of miserable and dejected life.(Malik Badri, 2013)
For endogenous depression al-Balkhi prescribed physical medical treatment (purifying the blood) rather than cognitive work, because the etiology was organic. For reactive depression, he recommended a structured program of cognitive strategies: persuasive counseling, reframing the loss against the broader certainty that all losses are eventually absorbed by time, and recognizing that prolonged depression harms the body more than the original loss.(Malik Badri, 2013) Badri argues that this endogenous/reactive distinction — credited in Western historiography to Kraepelin at the end of the nineteenth century — was made by al-Balkhi ten centuries earlier.(Malik Badri, 2013) Whether or not this claim of priority holds in all its details, al-Balkhi’s work demonstrates that a conceptually sophisticated account of depressive illness existed outside the Galenic tradition at least as early as the ninth century.
Islamic medicine also produced substantive accounts of melancholy in the Galenic mold.(Ullmann, 1978) The transmission of Galenic ideas ran through Rufus of Ephesus, who influenced Galen and whose views then passed to ninth-century Arab physicians; Ishaq ibn Imran in particular drew on Rufus, and his work in turn reached Constantinus Africanus (d. 1087), whose Latin translation De Melancholia was read widely through the Middle Ages and Renaissance.(Lawlor, 2012) Ishaq defined melancholy as a somatic illness caused by black bile vapour rising to the seat of reason, producing “a feeling of dejection and isolation because of something the patients think is real but which is in fact unreal.”(Ullmann, 1978) Ishaq classified melancholy into three types: idiopathic (originating in the brain itself), sympathetic (black bile rising from the whole body), and hypochondriac (black bile affecting the stomach region and thence the brain).(Ullmann, 1978) Alongside this humoral tradition, al-Razi composed On Spiritual Medicine, arguing that the pursuit of pure knowledge and the avoidance of mental afflictions such as greed, lust, and fear was both a philosophical imperative and a path to bodily health, linking moral philosophy directly to medicine.(Pormann, 2007)
Renaissance and the Double Tradition
The Renaissance produced two competing traditions for understanding melancholy: the Galenic tradition treated it as a serious physical illness, and the Aristotelian/Ficinian tradition celebrated it as a vehicle for creative genius.(Lawlor, 2012) Lawlor observes that both traditions coexisted in Renaissance literary culture, “hopelessly entangled,” with no clear resolution between them.(Lawlor, 2012)
Marsilio Ficino (1433–99), himself a melancholic, elaborated the Aristotelian idea of melancholic genius in De Vita Libri Tres (1482–9) and connected it to Neoplatonic astrology, identifying Saturn as the planet of scholars and men of genius.(Lawlor, 2012) Robert Burton (1577–1640), a solitary scholar at Oxford, wrote the most encyclopaedic Renaissance treatment of melancholy, the Anatomy of Melancholy, and was himself a melancholic whose sedentary, academic lifestyle was thought to cause humoral imbalance.(Lawlor, 2012)
The “dark night of the soul” tradition within Christian mysticism offered a third framework: depression as a necessary stage of purification on the soul’s journey toward divine knowledge. For mystics like John of the Cross, a degree of spiritual depression was not something to eliminate but something to pass through, evidence of genuine engagement with the inner life.(Lawlor, 2012)
The Early Modern Dissolution of Melancholy
By the early eighteenth century, the old humoral vocabulary was losing credibility under the pressure of the New Science, and a terminological proliferation replaced it. George Cheyne noted that before the time of Queen Anne, “melancholy, as the name for morbid depression, had been largely replaced by hypochondria, spleen, hysteria, and vapours, all four terms denoting the same disorder.”(Lawlor, 2012) Thomas Willis, working within the iatrochemical tradition, rejected the melancholic humor itself and attributed depression instead to a malfunction of the animal spirits, which in melancholics became “obscure, thick, and dark.”(Lawlor, 2012)
Cheyne’s own contribution was to frame depression as a disease of English civilization — the “English Malady” of his 1733 book — caused by the nation’s commercial prosperity, dietary excess, and sedentary luxury. “Since our wealth has increased and our navigation has been extended,” he wrote, “we have ransacked all the parts of the globe to bring together its whole stock of materials for riot, luxury, and to provoke excess.”(Lawlor, 2012) Depression as a symptom of modern life — an argument that would recur in neurasthenia in the nineteenth century and in burnout discourse in the twenty-first — was already fully formed in 1733.
Roy Porter observed that the Enlightenment’s detachment of mental disorder from divine or diabolic causation removed the moral stigma of nervous disorder for the elite, allowing its “gentrification” into markers of refined sensibility and artistic genius.(Lawlor, 2012)
The moral dimensions of depressive suffering had a longer history still. Acedia arose among fourth-century Egyptian desert monks as nostalgia for former life, low mood, ennui, and generalized misery, framed entirely within a moral-theological register rather than a medical one.(Lawlor, 2012) In the medieval period, acedia and Galenic melancholia coexisted with some symptomatic overlap, but because melancholia possessed a physical basis in humoral theory, it was regarded as less morally stigmatizing than acedia, which remained a failure of the will and spirit.(Lawlor, 2012)
The Nineteenth Century: Melancholia Becomes Affective
Victorian psychiatry accomplished a conceptual shift that made modern depression possible: it moved melancholy from a disorder of the intellect to a disorder of the emotions (what the century called the “affective” faculties). Lawlor identifies this shift as partly enabled by the professionalization of psychology, by advances in brain and nerve anatomy, and by the overall drift from defining melancholy through its delusional content to defining it through its mood.(Lawlor, 2012)
Esquirol, following Pinel, proposed replacing the term “melancholia” with “lypemania” — a cerebral malady characterized not primarily by delusions but by debilitating sadness as a defining criterion.(Lawlor, 2012) Berrios marks this as the first formal linkage of the diagnostic category to sustained depressed affect.(German E. Berrios & Roy Porter (eds.), 1995) Esquirol’s move narrowed the field: it excluded from melancholia the states with exalted mood and confined the diagnosis to the lowering disturbances.
In 1854, Jules Baillarger and Jean-Pierre Falret independently described the alternation of melancholic and manic episodes — Baillarger calling it folie à double forme, Falret calling it folie circulaire — establishing what would become the bipolar concept and separating episodes of depression from episodes of mania as components of a single cyclical disease.(Lawlor, 2012) The priority dispute between them was a genuine simultaneous discovery: both described the alternation of mania and melancholia, with Falret emphasizing the interval of lucidity between episodes.(German E. Berrios & Roy Porter (eds.), 1995) Aretaeus of Cappadocia had anticipated something like this relationship between mania and melancholia nearly eighteen centuries earlier, but the systematic nosological formulation was new.(German E. Berrios & Roy Porter (eds.), 1995)
Not all late-nineteenth-century clinical work pointed toward pessimism about prognosis. Dreyfus (1907) demonstrated that patients diagnosed with involutional melancholia had a better outcome than the prevailing view allowed, challenging the assumption that this age-related depressive state was a distinct degenerative disease with an unfavorable course.(German E. Berrios & Roy Porter (eds.), 1995) This finding contributed to the pressure on nosologists to reassess whether involutional melancholia deserved a separate category at all.
Within this period of conceptual consolidation, the Scottish asylum physician Thomas Clouston offered a characteristic Victorian synthesis. His 1883 definition presented melancholia as a clinical entity encompassing emotional depression, loss of self-control, delusions, suicidal impulses, and bodily symptoms — a formulation that tied the diagnosis tightly to observable behavioral and somatic features rather than to humoral theory.(Jansson, 2021) Wilhelm Griesinger, whose Mental Pathology and Therapeutics (1867) supplied the refrain for the “brain psychiatry” of the second half of the century, argued that all mental disease was a symptom of brain malfunction; his concept of unitary psychosis treated each mental illness as a different stage on a single continuum, and he described melancholia as the first, most fundamental move toward more serious states.(Lawlor, 2012) After the eighteenth century, melancholy became progressively feminized: the condition was increasingly coded as female, reflecting broader cultural anxieties about women’s bodies, reproduction, and social roles rather than any demonstrable change in clinical presentation.(German E. Berrios & Roy Porter (eds.), 1995)
Kraepelin’s synthesis in the eighth edition of his Lehrbuch (1913) cut through the proliferating classifications of the late nineteenth century by creating an overinclusive category of manic-depressive insanity that absorbed most previous melancholias.(German E. Berrios & Roy Porter (eds.), 1995) His big division — between deteriorating disorders (dementia praecox, now schizophrenia) and episodic affective disorders (manic depression) — structured psychiatric classification for the rest of the century.(Lawlor, 2012) Kraepelin attributed up to eighty percent of manic-depressive cases to hereditary predisposition and placed heavy emphasis on internal rather than external factors.(Lawlor, 2012) Jansson identifies Kraepelin’s 1899 nosology as the key transitional moment between the Victorian biomedical category of melancholia and the twentieth-century concept of depression. Importantly, existing histories of this transition have overlooked two crucial drivers of the shift: the appropriation of physiological language to describe emotion as a physiological event, and the role of asylum statistics in the development of diagnostic categories.(Jansson, 2021)
The word “depression” itself entered medical dictionaries by the 1860s, borrowed from cardiovascular medicine where it denoted a reduction in functional activity. Its psychiatric application initially meant a lowering of mental function generally, not specifically sadness; the emotional meaning came later.(German E. Berrios & Roy Porter (eds.), 1995) Berrios’s broader argument is that the affective disorders as a category emerged from the asynchronous convergence of three separate historical elements — words, concepts, and behaviors — and that this convergence only completed in the early twentieth century.(German E. Berrios & Roy Porter (eds.), 1995)
The Psychoanalytic Interlude
Freud’s 1917 essay “Mourning and Melancholia” distinguished normal mourning (conscious grief with a cause) from pathological melancholia (unconscious anger directed against the introjected love object), defining melancholia as a response to loss rooted in childhood.(Lawlor, 2012) Karl Abraham provided the first psychoanalytic account of depression, identifying it as anger directed against the self rather than outward, and linking both anxiety and depression to the repression of unconscious conflicts.(Lawlor, 2012)
Melanie Klein’s “depressive position” (first formulated 1935) deepened the developmental account: the infant’s experience of weaning involves intense separation and loss, creating a depressive template to which adult losses can later return. Events in adulthood might trigger those original infantile depressive feelings through object-relations dynamics.(Lawlor, 2012)
Freud’s theoretical contribution served a further organizational function in this history. His work marked the conceptual boundary between neurosis — functional, later identified with depression — and psychosis — organic, later identified with mania. This Freudian demarcation helped shape the twentieth-century unipolar/bipolar distinction.(German E. Berrios & Roy Porter (eds.), 1995)
Adolf Meyer proposed in 1905 that “melancholia” should be replaced entirely by “depression” as a more neutral term that would not imply theoretical commitments the profession could not yet justify.(Lawlor, 2012) He was “desirous of eliminating the term melancholia, which implied a knowledge of something that we did not possess.” Meyer’s proposed terminology — depression as a pragmatic placeholder — eventually prevailed, but not in the way he intended: it was pressed into service as a precise diagnostic category in 1980.
DSM-III and the Construction of Major Depressive Disorder
The US–UK Diagnostic Project of 1972 exposed the alarming lack of diagnostic consensus that prompted the DSM-III’s symptom-based overhaul, as British psychiatrists diagnosed depression five times as often as Americans.(Lawlor, 2012) David Rosenhan’s 1973 Science paper further damaged psychiatry’s scientific credibility: pseudo-patients feigning hearing voices were admitted, diagnosed as schizophrenic despite normal behavior, and later discharged as “schizophrenic in remission,” leading legal scholars to mock psychiatry’s claims to competence.(Andrew Scull, 2015)
The Feighner criteria (1972), developed at Washington University in St. Louis, established the first explicit symptom checklist for depression: dysphoric mood, five of eight additional symptoms, one month duration, no other illness explaining the presentation.(Lawlor, 2012) Robert Spitzer’s Research Diagnostic Criteria (1978) lowered the time threshold from one month to two weeks without explanation, and permitted loss of interest to substitute for dysphoric mood, reducing the diagnostic bar further.(Lawlor, 2012)
DSM-III (1980) adopted this symptom-based framework and published it as the official diagnostic system of American psychiatry, thereby breaking with both Freudian and Meyerian traditions and providing what Lawlor calls the “bible of the New Depression.”(Lawlor, 2012) The framework was explicitly atheoretical — it made no claims about causes and demanded no specific etiology — and was designed primarily to improve inter-rater reliability. The motivation was in part historical: psychiatry had become averse to etiological nosology after its earlier dependence on Freudian theory, and the DSM-III’s symptom-based diagnoses improved inter-rater reliability while producing, as a consequence, poor diagnostic validity.(Stegenga, 2018) Scull notes that the DSM-III task force under Spitzer abandoned any pretense at causal validity: questions about whether the new classificatory system corresponded to aetiological distinctions that made biological sense “were simply set to one side.”(Andrew Scull, 2015)
The consequences were significant. Because DSM-III’s Major Depressive Disorder required only five of nine possible criteria, two patients sharing only one symptom (say, fatigue) could receive the same diagnosis.(Stegenga, 2018) This heterogeneity problem suggests that “depression” as defined by DSM-III is not a single disease but a family of conditions with potentially different etiologies grouped by symptom overlap. The removal of the bereavement exclusion in DSM-5 extended the diagnosis to people grieving the loss of a spouse within weeks of the death, which critics argue pathologizes normal grief.(Stegenga, 2018)
Lawlor argues that the DSM-III framework produced “a massive pathologization of normal sadness” that, ironically, made depressive diagnosis less rather than more scientifically valid — extending powerful psychoactive medications into a population much of whose suffering reflected normal responses to divorce, illness, financial ruin, and social isolation rather than biological dysfunction.(Lawlor, 2012) The gender bias built into the diagnostic criteria compounded this: Feighner’s study samples included twice as many women as men, and symptoms like irritability that might have drawn in more male presentations were excluded from the criteria.(Lawlor, 2012)
The Pharmaceutical Revolution and Its Discontents
The mass expansion of depression as a diagnosis ran alongside and was partially driven by the development of effective antidepressant drugs. Pharmaceutical company Geigy had refused to fund the development of imipramine in the mid-1950s because it judged the depression market to be insignificant.(Lawlor, 2012) The contrast with Prozac’s cultural ascendancy four decades later is striking and suggests that the disease, at least in its mass-market form, partly followed the drug.
The chemical imbalance theory — the claim that depression is caused by deficiency in serotonin or norepinephrine — provided the explanatory framework for antidepressant prescribing, but the evidence for it was always weaker than its popularity suggested. Lawlor notes that only roughly a quarter of depressives show low serotonin, and the theory cannot explain why SSRIs alter serotonin levels immediately but take weeks to affect mood.(Lawlor, 2012) Irving Kirsch’s meta-analysis of clinical trial data, obtained through Freedom of Information requests and including unpublished trials that pharmaceutical companies had not chosen to submit to regulators, found antidepressants to be only marginally more effective than placebos even in severe depression.(Lawlor, 2012) David Healy had argued earlier that the rise of depression as a disease category was driven by the introduction of psychiatric drugs in the 1950s hospital context — that pharmaceutical development shaped the nosology rather than merely responding to it.(Lawlor, 2012)
Prozac (fluoxetine) became the second best-selling drug in the world by 1994, its cultural ascendancy fueled by Peter Kramer’s Listening to Prozac (1993) and Elizabeth Wurtzel’s Prozac Nation (1994), which confirmed depression’s public image as a primarily female disorder and suggested that SSRIs could do more than treat illness — they could improve baseline personality.(Lawlor, 2012) By 2010, global antidepressant sales reached $20 billion annually.(Andrew Scull, 2015)
Depression Across Cultures
The history traced above is overwhelmingly Western. Seen from cross-cultural psychiatry, the category of depression as defined by DSM-III is not merely contested — it may constitute what Arthur Kleinman called a category fallacy when projected onto populations outside the cultural context in which it was developed. The case of neurasthenia in China is the clearest illustration. George M. Beard coined the term “neurasthenia” in the late nineteenth century as exhaustion of the nervous system, attributing it specifically to the pressures of modern American civilization — a framing that cast the disorder as both organic and culturally produced.(Lawlor, 2012) The label served as a standard diagnosis for decades before falling out of favor in American psychiatry. The American Psychiatric Association removed it from DSM-III in 1980. Yet in the same year, neurasthenia remained an official diagnosis in China and in the WHO’s ICD-9 — the same cluster of symptoms was differently named across professional taxonomic systems, each reflecting local assumptions about what counted as a real disease.(Arthur Kleinman, 1988)
Kleinman’s study of one hundred neurasthenic patients at the Hunan Medical College outpatient psychiatry clinic demonstrated the problem concretely. Using a standard North American psychiatric protocol and DSM-III criteria, he showed that most of these patients could be rediagnosed as cases of major depressive disorder. But the rediagnosis proved clinically incomplete: unlike most chronically depressed patients, the Chinese neurasthenics responded only partially to antidepressant medication. Their chief somatic complaints and medical help-seeking ended only when they resolved underlying work and family problems — suggesting that the biomedical category captured part of the phenomenon but missed the culturally shaped illness behavior overlying it.(Arthur Kleinman, 1988)
Kleinman’s broader argument was that what is now called depression has occupied different categories in different cultural and historical moments: medical disorder (humoral imbalance), religious problem (guilt or sinfulness), moral weakness (acedia), or fate. In traditional Chinese medicine, only madness and hysteria counted as mental disorders; states that Western psychiatry would now categorize as depression were interpreted as medical disorders or as trouble caused by gods, ghosts, and ancestors.(Arthur Kleinman, 1988) When Western diagnostic instruments are translated and administered cross-culturally, they may produce reliable findings — consistent observations between trained clinicians — without producing valid ones, because the categories being applied lack coherence in the local context.(Arthur Kleinman, 1988)
Kleinman argued that dysthymic disorder, applied globally, may amount to the medicalization of realistic responses to chronic deprivation: in much of the world, endemic hopelessness and helplessness are responses to actual conditions of persistent loss, and powerlessness is not a cognitive distortion but an accurate perception of one’s place in an oppressive social system.(Arthur Kleinman, 1988)
The Stress-Biology of Depression
A physiological account of depression running parallel to the historical and nosological one emerges from stress neurobiology, and it supplies mechanisms that the DSM’s symptom-list framework deliberately set aside.
Anhedonia as the core
Robert Sapolsky’s account of depression opens with its defining feature: anhedonia, the inability to feel pleasure. If forced to a single sentence, he describes depression as “a genetic/neurochemical disorder requiring a strong environmental trigger whose characteristic manifestation is an inability to appreciate sunsets.”(Sapolsky, Robert M., 2004) Stressing a laboratory rat produces measurable anhedonia: it takes a stronger electrical current than usual in the rat’s pleasure pathways to activate a sense of pleasure, raising the threshold for experiencing reward. Glucocorticoids alone can produce the same elevation in pleasure threshold, placing the adrenal axis at the proximal mechanism for at least this cardinal feature of depression.(Sapolsky, Robert M., 2004)
The glucocorticoid hypothesis and feedback resistance
Many depressed patients show elevated circulating glucocorticoids — not because of excessive stress signals from outside, but because the brain’s negative feedback system fails. Normally, rising glucocorticoid levels are detected by the brain, which then shuts down CRH secretion. In depression, this feedback regulation breaks down: the brain becomes less effective at sensing and responding to elevated glucocorticoid levels, driving secretion higher still.(Sapolsky, Robert M., 2004) This “feedback resistance” provides a direct therapeutic prediction: drugs that lower glucocorticoid secretion should have antidepressant effects in the subgroup of depressives whose depression is glucocorticoid-mediated, and several solid clinical demonstrations have borne this out.(Sapolsky, Robert M., 2004)
The catecholamine/glucocorticoid distinction illuminates the relationship between anxiety and depression: catecholamine excess characterizes still-trying-to-cope anxiety states, while glucocorticoid excess characterizes giving-up depressive states — demonstrable in rat models where aversive conditions initially produce anxious catecholamine-dominated behavior that transitions, if sustained, to passive glucocorticoid-dominated depression.(Sapolsky, Robert M., 2004) Moderate transient stress produces dopamine release in pleasure pathways and feels like stimulation; severe or prolonged stress produces dopamine depletion and dysphoria, mechanistically connecting the pleasant end of the stress spectrum to the disease state at the chronic end.(Sapolsky, Robert M., 2004)
Hippocampal damage and learned helplessness
Depression is associated with hippocampal atrophy that emerges during the illness and is proportional to its duration; volume loss persists years to decades after remission, suggesting irreversible structural change rather than a transient state.(Sapolsky, Robert M., 2004) Human evidence links excess glucocorticoids to hippocampal volume loss across six clinical conditions, including major depression.(Sapolsky, Robert M., 2004) Unlike Cushing’s syndrome — where volume recovers when the tumor is removed and glucocorticoids normalize — hippocampal volume loss in major depression appears largely permanent, persisting years to decades after medication-controlled remission.(Sapolsky, Robert M., 2004)
Martin Seligman’s learned helplessness model provides an animal framework for this biology. Animals subjected to repeated uncontrollable stressors develop a syndrome that replicates key features of human depression: anhedonia, motivational failure, cognitive distortion, elevated glucocorticoids, norepinephrine depletion, and disrupted sleep architecture. The model implies that the neurobiological damage of depression derives not merely from stress per se but specifically from the loss of control over stressor exposure.(Sapolsky, Robert M., 2004)
The serotonin transporter gene and gene-environment interaction
The “chemical imbalance” theory — depression as simple serotonin deficiency — is an oversimplification that cannot explain the weeks-long delay between SSRI administration and mood improvement. A more adequate picture emerges from gene-environment interaction. The serotonin transporter gene (5-HTT), which codes for the reuptake pump that SSRIs inhibit, comes in allelic variants that differ in their efficiency at clearing serotonin from the synapse. Crucially, glucocorticoids regulate how much 5-HTT protein the gene produces, creating a direct link between the HPA axis and serotonin availability.(Sapolsky, Robert M., 2004) Carrying a lower-efficiency allele of this gene does not by itself predict depression; the variant increases risk only when coupled with a history of repeated major stressors.(Sapolsky, Robert M., 2004)
Sex differences and reproductive transitions
Depression risk is substantially higher in women than men for unipolar depression, with particular vulnerability at reproductive transitions — menstruation, postpartum period, and menopause — linked to the large fluctuations in estrogen and progesterone that occur at these points and their downstream effects on neurotransmitter metabolism.(Sapolsky, Robert M., 2004)
Kindling: from reactive to autonomous depression
Stressors function as triggers for the first few depressive episodes; after the third or fourth episode, however, the disease takes on a self-sustaining rhythm largely independent of external stressors. The mechanism is not yet understood, but the clinical pattern suggests a progressive neurobiological kindling in which each depressive episode lowers the threshold for the next, eventually producing a disorder that runs on its own internal timetable regardless of what is happening in the person’s life.(Sapolsky, Robert M., 2004)
The immune-mood axis
Chronic immune activation is more likely to cause depression than equivalently severe illnesses not involving the immune system. The mechanism runs through cytokines: these immune-signaling molecules can cross into the brain and interact directly with norepinephrine, dopamine, and serotonin systems, establishing an immune-mood axis that connects autoimmune disease, chronic infection, and inflammatory states to depressive illness.(Sapolsky, Robert M., 2004)
The Freudian contribution reread
Out of the Freudian tradition came one of the more clinically accurate descriptions of the phenomenology: depression as “aggression turned inward.” This formulation accounts for the loss of pleasure, the psychomotor retardation, suicidal impulses, and — notably — the elevated glucocorticoid levels, which are characteristic of the internal conflict state rather than external-threat states. From this angle the depressed patient is not simply too lethargic to function but is exhausted by what may be the most draining emotional conflict of a human life, conducted entirely internally.(Sapolsky, Robert M., 2004)
Whitaker’s Critique: Antidepressants and the Disability Epidemic
The investigative journalist Robert Whitaker, writing in Anatomy of an Epidemic (2010), assembled a body of research to argue that the era of mass antidepressant prescribing has produced more chronic disability than it has relieved. Whitaker is not a clinical researcher; his work draws on published studies rather than original data, and his interpretations are disputed by mainstream psychiatry. His evidence is documented here as a counterpoint grounded in published research, not as authoritative conclusion.
Whitaker begins with a baseline contrast that shapes his entire argument. Community surveys from the 1930s and 1940s found fewer than 1 in 1,000 adults suffered from clinical depression in any given year. In 1955 — before antidepressants existed as a drug class — only 38,200 depressed patients occupied the nation’s psychiatric hospitals, a disability rate of 1 in 4,345.(Whitaker, Robert, 2010) Kraepelin’s own follow-up data on depressed patients supported a similar picture: 60 percent of his patients with depression-only diagnoses had a single episode; 50 percent never recurred; only 1 in 10 became chronically ill.(Whitaker, Robert, 2010) The NIMH and academic psychiatrists of the 1960s and 1970s held a consensual view that “most depressions terminate in spontaneous remissions” and “will run their course and terminate with virtually complete recovery without specific intervention.”(Whitaker, Robert, 2010) This was the baseline against which the drug era’s outcomes must be measured.
The case of Melissa Sances — diagnosed at 16, still drawing SSDI disability payments years later — illustrates the transformation Whitaker documents: a condition that in the 1950s was episodic and largely self-resolving had become, for many patients, a chronic disability identity defined by pharmaceutical management.(Whitaker, Robert, 2010)
Whitaker notes that from 1987 to 2007, the number of adults disabled by affective disorders (depression and bipolar illness) grew to an estimated 1.4 million, driven by a surge that began in the 1990s — precisely the decades in which antidepressant prescribing expanded most rapidly.(Whitaker, Robert, 2010) The paradox he identifies is that psychiatric medications are widely accepted as effective, yet the number of disabled mentally ill has risen dramatically since 1955, accelerating in the two decades of explosive prescribing after Prozac’s approval in 1987.(Whitaker, Robert, 2010)
The serotonin hypothesis of depression — the claim that the condition reflects a deficit of serotonin in the brain — was constructed from drug observations rather than disease pathology. Whitaker traces it to Bernard Brodie’s 1955 observation that reserpine depleted serotonin and made rabbits lethargic, while iproniazid and imipramine blocked this depletion.(Whitaker, Robert, 2010) Joseph Schildkraut’s 1965 “catecholamine hypothesis,” which formalized this into a theory, was acknowledged by its own author as “at best a reductionistic oversimplification of a very complex biological state.”(Whitaker, Robert, 2010)
The research testing the serotonin hypothesis in humans failed to confirm it. Multiple independent studies from 1969 to 1974 repeatedly failed to find significantly lower serotonin metabolites (5-HIAA) in depressed patients versus controls; a 1974 Bowers study found unmedicated depressed patients had perfectly normal 5-HIAA levels.(Whitaker, Robert, 2010) Mendels and Frazer (1974) found that reserpine — the drug used to infer low-monoamine causation of depression — induced depression in only 6 percent of hypertensive patients and actually lifted spirits in some depressed patients when used as treatment in England in 1955.(Whitaker, Robert, 2010) The 1984 NIMH study found no relationship between low-serotonin subgroup status and response to amitriptyline, directly refuting the prediction that the drug would be most effective in “low-serotonin” depressed patients.(Whitaker, Robert, 2010) By 2003–2005, leading psychiatrists including David Burns (Stanford) and Kenneth Kendler were publicly stating that no convincing biological deficit had ever been found for depression or other psychiatric disorders, despite ongoing pharmaceutical advertising of the serotonin theory.(Whitaker, Robert, 2010)
The efficacy case for antidepressants is weaker than standard clinical narratives suggest. A 1969 NIMH review of the original clinical trial literature found that the more rigorously controlled a study, the lower the observed improvement rate — the net drug-over-placebo benefit was only 15 percentage points (61 percent drug response versus 46 percent placebo), leading the reviewers to conclude that “differences…not impressive.”(Whitaker, Robert, 2010) Irving Kirsch’s meta-analysis of published and unpublished trial data — including trials the pharmaceutical companies had not submitted to regulators — found the drug-placebo difference averaged only 1.8 points on the Hamilton Depression Rating Scale, well below the 3-point threshold the FDA itself uses to define clinical significance; Kirsch concluded that antidepressants should be prescribed only for the most severely depressed patients.(Whitaker, Robert, 2010) A three-way comparison of St. John’s wort, Zoloft, and placebo found response rates of 24 percent, 25 percent, and 32 percent respectively — both the drug and the herb failed to beat placebo, with placebo performing best.(Whitaker, Robert, 2010)
A patient Whitaker cites captures the resulting confusion for someone who had been told her depression was a neurotransmitter problem: after years on Zoloft, then Prozac, then Effexor, she wondered what might have happened had she been helped instead with diet and exercise — whether there had been a “neurotransmitter problem” at all, or whether the story she had been given shaped both her self-understanding and her treatment course.(Whitaker, Robert, 2010)
The longer-term evidence Whitaker assembles concerns antidepressant-induced chronicity. European psychiatrists raised these concerns earlier than American researchers. Observers in the late 1960s and 1970s documented that patients treated with antidepressants appeared to have “shortening intervals” between episodes, a process some called “chronification” — a change to a more chronic course directly associated with drug treatment.(Whitaker, Robert, 2010) In female patients specifically, more systematic antidepressant medication was associated with both increased recurrence rates and decreased cycle duration — the drugs were correlating with a more aggressive disease course, not a milder one.(Whitaker, Robert, 2010) Studies from 1973 to 1984 found that 50 to 71 percent of patients relapsed within 6 to 18 months of antidepressant withdrawal; Ross Baldessarini’s analysis found 50 percent relapse within 14 months after discontinuation, and higher relapse rates for patients who had been treated longer — a pattern consistent with drug-induced sensitization rather than underlying disease severity.(Whitaker, Robert, 2010)
Antidepressant exposure also appears to accelerate diagnostic transition to bipolar disorder. Whitaker documents that antidepressants convert depressed patients to bipolar illness at a rate of 7.7 percent per year — three times the rate for unexposed patients; estimates suggest 20 to 40 percent of unipolar depressed patients will eventually convert to bipolar; and in surveys of bipolar association membership, 60 percent of members reported that their bipolar disorder began only after antidepressant exposure.(Whitaker, Robert, 2010) If these figures reflect causation rather than correlation, the mass prescribing of antidepressants for unipolar depression would be a significant contributor to the subsequent expansion of bipolar diagnoses.
The Italian researcher Giovanni Fava’s 1994 editorial proposed that psychotropic drugs worsen the long-term course of the illnesses they treat through a shared biological mechanism: perturbation of neurotransmitter systems leads to compensatory processes that “operate unopposed” when drug treatment ends, producing withdrawal symptoms and increased vulnerability to relapse.(Whitaker, Robert, 2010) Fava argued that antidepressants may cause “irreversible receptor modifications” that sensitize the brain to depression, potentially “propelling the illness to a more malignant and treatment-unresponsive course.”(Whitaker, Robert, 2010) When Fava raised these concerns, Donald Klein from Columbia University told Psychiatric News the issue would not be investigated: “The industry is not interested, the NIMH is not interested, and the FDA is not interested. Nobody is interested.”(Whitaker, Robert, 2010)
Whitaker documents that psychiatry responded to concerns about antidepressant-induced chronicity not by investigating them but by reconceptualizing depression as inherently chronic and recurrent — rewriting textbooks to state the old evidence of good outcomes was “flawed” and that depression is “a highly recurrent and pernicious disorder” — thereby providing a rationale for indefinite drug maintenance.(Whitaker, Robert, 2010) The clinical trials that followed tested this new paradigm directly. John Rush’s 2004 real-world trial of 118 patients given antidepressants with maximum clinical support found only 26 percent responded and only 6 percent achieved sustained remission over one year. The NIMH STAR*D trial of 4,041 patients found fewer than 20 percent remitted and stayed well for one year.(Whitaker, Robert, 2010)
Several naturalistic outcome studies Whitaker cites found unmedicated depression had better short- and long-term outcomes than medicated. British patients’ symptoms decreased 62 percent versus 33 percent in six months for medicated patients; Dutch patients showed 76 percent versus 50 percent recovery without relapse at ten years; Canadian patients spent 11 versus 19 weeks per year depressed.(Whitaker, Robert, 2010) A WHO study in fifteen cities found that the 484 depressed patients not exposed to psychotropics had the best twelve-month outcomes in “general health,” mildest depressive symptoms, and lowest percentage “mentally ill” — while antidepressant-treated patients had the highest rates of “continued depression.”(Whitaker, Robert, 2010) A Canadian disability study of 1,281 employees found that unmedicated depressed workers returned to work in 77 days versus 105 days for medicated, and only 9 percent went on to long-term disability versus 19 percent of those who took an antidepressant.(Whitaker, Robert, 2010) Michael Posternak’s 2006 study of 84 unmedicated patients relapsing after a first episode found 85 percent recovered within one year without medication — confirming Kraepelin’s estimate that untreated depression usually clears within six to eight months — and concluded “it would be extremely difficult for any intervention to demonstrate a superior result to this.”(Whitaker, Robert, 2010)
Whitaker argues these studies collectively suggest that antidepressants transform an episodic condition into a chronic one for many patients. The per-capita disability rate from depression rose from 1 in 4,345 in 1955 to a condition now affecting 15 million Americans annually, 58 percent of them “severely impaired.”(Whitaker, Robert, 2010) SSRIs cause a multitude of troubling side effects including sexual dysfunction, REM sleep suppression, muscle tics, fatigue, emotional blunting, and apathy; long-term use is associated with memory impairment, problem-solving difficulties, loss of creativity, and learning deficiencies — side effects acknowledged by Maurizio Fava as “quite common” and receiving insufficient research attention.(Whitaker, Robert, 2010)
Whitaker extends his critique beyond antidepressants to the broader class of psychotropics frequently co-prescribed with them. A 2007 French survey of 4,425 long-term benzodiazepine users found that 75 percent described themselves as “markedly ill to extremely ill.” A 2004 Australian systematic review of long-term benzodiazepine use found that long-term users were “consistently more impaired than controls across all cognitive categories” — making chronic benzo use, commonly layered onto antidepressant regimens, a further contributor to functional deterioration in depressed patients.(Whitaker, Robert, 2010)
Whitaker also documents exercise as a comparator. James Blumenthal’s Duke University study found that exercise alone produced an 8 percent relapse rate in depressed patients over ten months, compared to a 38 percent relapse rate for Zoloft alone and a 31 percent relapse rate for Zoloft plus exercise — making exercise the most effective long-term treatment of the three.(Whitaker, Robert, 2010) The UK’s National Health Service began prescribing exercise for depression after reviewing evidence showing 70 percent response rates and better long-term outcomes than antidepressants.(Whitaker, Robert, 2010) Adding Zoloft to exercise provided no benefit over exercise alone in Blumenthal’s study, which Whitaker interprets as evidence that the drug’s biological action was either irrelevant or actively counterproductive compared to the benefits of behavioral change.(Whitaker, Robert, 2010)
Mainstream psychiatry disputes Whitaker’s reading of this evidence on several grounds: the naturalistic studies confound selection (patients who can remain unmedicated may have milder illness), the chronicity data reflect worsening illness rather than drug effects, and short-term symptom relief justifies treatment even when long-term outcomes are uncertain. The underlying questions — whether antidepressants help or harm patients over the long term, and whether disability expansion tracks drug prescribing causally — remain genuinely contested and under-investigated.
Contested Boundaries
The category of depression has been contested from multiple directions. Within psychiatry, Max Fink and Michael Taylor argued in 2003 for the restoration of melancholia as a distinct diagnostic category separate from Major Depressive Disorder, claiming that genuine melancholia — with its characteristic psychomotor features, early-morning awakening, diurnal variation, and responsiveness to ECT — is biologically and clinically distinguishable from the broader depression category into which it has been dissolved. Aubrey Lewis’s 1934 Maudsley Hospital study had already challenged the endogenous/reactive distinction by showing that most ostensibly endogenous depressions had external precipitants.(Lawlor, 2012)
From outside the biological paradigm, critics have challenged whether the category adequately distinguishes biological illness from normal suffering. The question is not trivial: Jerome Wakefield’s harmful dysfunction analysis holds that many presentations currently diagnosed as depression fail to constitute genuine disorders because the emotional response, while distressing, may be a normal response to genuinely bad circumstances rather than a failure of a normally functioning biological mechanism.(Unknown, unknown) Stegenga observes that the heterogeneity problem — two patients qualifying for the same diagnosis while sharing only a single symptom — raises genuine questions about whether “depression” names one disease or several.(Stegenga, 2018)
The cultural variation in how depressive suffering is expressed, framed, and presented further complicates the picture. Jansson’s argument that Victorian melancholia was not a transhistorical condition but a historically-specific biomedical object created through physiological language and asylum statistics applies equally to its successor: “depression” as it appears in the DSM is itself a historically-specific construction, not a direct window onto universal human suffering.(Jansson, 2021)
Human Notes
See Also
- melancholia
- manic-depressive-insanity
- dsm
- neurasthenia
- psychoanalysis
- anti-psychiatry
- medicalization
- serotonin-hypothesis
- abu-zayd-al-balkhi
- adolf-meyer
- thomas-szasz
Sources
Evidence cited from: berrios-porter-historyclinicalpsychiatry-1995, lawlor-from-melancholia-to-2012, jansson-from-melancholia-to-2021, badri-abuzaydalbalkhi-2013, stegenga-care-and-cure-2018, scull-madnesscivilization-2015, sep-philosophy-medicine, kleinman-rethinkingpsychiatry-1988