Experimental Method in Medicine

Summary

The experimental method in medicine is the practice of deliberately testing ideas about the body by manipulating conditions and observing what follows — and then using those observations to confirm, revise, or reject the ideas that prompted the test. It is older than the word “experiment” and has never had a single form. Galen opened living animals in the second century CE and drew conclusions about nerves; Harvey counted heartbeats and calculated blood volumes in the seventeenth century; Bernard trained himself to abandon a hypothesis the moment an unexpected result appeared in the nineteenth century. What connects these practices is a shared insistence that the body speaks through what it does under controlled conditions, not through what authorities say about it. How those results should be documented, who is entitled to trust them, and what counts as adequate proof have all been contested at every stage of the method’s history.

Ancient Precedents: Galen’s Vivisection Program

Before the experimental method had a name, Galen of Pergamum (129–c. 210 CE) practiced something recognizable as its core operation: he arranged conditions, intervened in a living body, and recorded what happened. Rocca documents that Galen’s primary experimental technique for establishing the function of the brain’s ventricles was systematic vivisection — exposing the brain of a living animal, applying pressure or making incisions to each ventricle in sequence, and observing the resulting effects on sensation and motion.(Rocca, 2003) The results were organized. Rocca details that Galen found a consistent gradient of harm: opening the posterior (fourth) ventricle caused the greatest loss of sensation and motion; the middle ventricle produced intermediate effects; incision of either anterior ventricle produced the least harm.(Rocca, 2003)

Galen’s experimental design included what we would now call methodological controls. Rocca notes that temperature control was treated as essential to valid vivisection results: Galen recognized that thermal fluctuation in the exposed brain would confound observations about pneumatic function and took deliberate precautions to maintain stable conditions during experiments.(Rocca, 2003) He also deployed ligation — tying off vessels or nerves — as what he himself called “the best as well as the clearest method of deciding the source of the body’s activities.”(Rocca, 2003) His carotid ligation experiments in animals, in which he tied off both carotid arteries and observed that the animal continued to move and breathe normally for the rest of the day, were designed to settle a real question about whether the brain drew its essential substance from the heart.(Rocca, 2003)

The experimental program was coherent, but Rocca’s analysis identifies its limits. In his carotid ligation work, Galen offered two incompatible explanations across different texts: in one work, the retiform plexus served as a reservoir sufficient to sustain the brain after the carotids were tied; in another, the brain drew directly on air entering through the nostrils, bypassing the cardiac supply entirely. Rocca finds that these two accounts cannot be fully reconciled, and that Galen never produces a unified synthesis that resolves the tension.(Rocca, 2003) The experimental results were real; the theoretical framework imposed on them shifted between texts. Rocca concludes that it is to Galen’s credit that he permits a window into his uncertainties in physiological experimentation — he makes no convincing attempt to dress speculation as dogma in this domain.(Rocca, 2003)

Galen’s approach to observational knowledge also carried an explicit pedagogical commitment. Rocca notes that he repeatedly urged pupils to see structures for themselves rather than accepting his authority, working to root out errors in predecessors and students alike.(Rocca, 2003) His concept of autopsia — direct personal seeing — served double duty: epistemologically, it secured knowledge through the observer’s own gaze rather than reliance on texts; rhetorically, it generated a sense of wonder at nature’s design.(Rocca, 2003) Rocca also documents that Galen’s public anatomical demonstrations — most notably his exposure of the recurrent laryngeal and intercostal nerves in living animals — were deliberately staged as responses to philosophical disputes about the seat of the soul and the origins of voluntary motion, using experimental spectacle to settle questions that textual argument alone could not resolve.(Rocca, 2003) Observation and persuasion were intertwined from the beginning.

Galen’s formal experimental methodology, as Rocca reconstructs it, was grounded in Aristotelian first principles supplemented by his selective reception of the anatomical legacy of Herophilus, Erasistratus, and Marinus.(Rocca, 2003) Herophilus and Erasistratus had practiced dissection and a form of experimentation that expanded Aristotle’s biological methodology; their Alexandrian work devoted considerable research to the brain and nerves and anatomically revealed it as a structure worthy of investigation.(Rocca, 2003) Galen built on this, and what he built remained largely unchallenged in its anatomical account of the brain until Thomas Willis in the seventeenth century.(Rocca, 2003)

Against Galen’s experimental tradition stood the Empiricist medical sect, which Rocca documents as founded by Philinus of Cos in Alexandria in the third century BCE.(Rocca, 2003) The Empiricists denied the epistemological validity of anatomical dissection, holding that experience of disease and case histories provided all the information required for medical practice.(Rocca, 2003) Their objection had a specific edge: the very act of dissection, they argued, produces changes in the appearance of organs under investigation, so that dissection is an unreliable source of knowledge about the living body.(Rocca, 2003)

Measurement and Mechanism: Santorio, Harvey, and the Early Modern Turn

The experimental method’s second major moment came not from anatomy but from measurement. Sigerist documents that Santorio Santorio (1561–1636), inspired by “the new methodology of the exact sciences,” invented the first clinical thermometer — a graduated capillary tube placed in the patient’s mouth — to measure body temperature objectively rather than by touch.(Henry E. Sigerist, 1933) He also invented the pulsilogium, a pendulum whose thread length was adjusted until the bob swung synchronously with the patient’s pulse, providing an objective measure of pulse rate.(Henry E. Sigerist, 1933) And over thirty years of self-experimentation using a large balance — sitting in one scale, food and drink on the other — Santorio quantified insensible perspiration by weighing the body, its ingested food, and its excreted matter, tracing variations under the influence of air, water, food, drink, sleeping, and waking.(Henry E. Sigerist, 1933) The resulting 1614 text, Ars de statica medicina, recorded the upshot of these thirty years in aphorisms.

Sigerist reads Santorio as a methodological contemporary of Harvey, making the same move from qualitative to quantitative observation and thinking in mechanical terms.(Henry E. Sigerist, 1933) But Santorio chose a less arresting problem(Henry E. Sigerist, 1933), overreached his conclusions, and buried his original observations inside voluminous commentaries on Galen and Avicenna.(Henry E. Sigerist, 1933) His instruments were not adopted until long after his death.(Henry E. Sigerist, 1933)

Hall documents the argument’s structure: Harvey demonstrated by direct experiment that the heart receives and expels during each contraction a substantial quantity of blood, not a few drops; by calculation he proved that on the lowest estimate of ventricular volume change, all the blood in the body passed through the heart more than once in half an hour.(Hall, A. Rupert, 1954) Hall characterizes Harvey’s great merit not as the introduction of new facts but as integrating known facts into a new and comprehensive generalization, by treating the vascular system as a hydraulic problem: the heart as a pump, veins and arteries as pipes, valves as mechanical valves, and blood as a fluid.(Hall, A. Rupert, 1954) Hall notes that Descartes’ justification of experimental inquiry in biology was of permanent value, as it treated physiological processes as subject to material causes rather than vital spirits or occult agencies.(Hall, A. Rupert, 1954)

Robert Boyle later recorded that Harvey said the venous valves first induced him to think of circulation — he reasoned that nature would not have placed so many valves without design, and no design seemed more probable than that blood was sent through arteries and returned through veins.(Hall, A. Rupert, 1954) The reasoning was teleological even as the proof was experimental. French confirms that Harvey’s natural philosophy was fundamentally experimental: his first discovery, the forceful systole, was found experimentally and had to be defended experimentally, and for the presentation of circulation in De motu cordis he relied on the force of direct experimental demonstration.(French, 1994)

Dear, analyzing the epistemology of Harvey’s method, notes that Harvey himself wanted experimental physiology to achieve certainty analogous to mathematics, citing geometry as his model — “geometry is a reasonable demonstration about sensibles from non-sensibles.”(Peter Dear, 2001) Harvey also legitimated his claims institutionally, informing potential critics that the learned Doctors of the College of Physicians had attended his ocular demonstrations and “been in full agreement” — using institutional authority as a bulwark against criticism.(Peter Dear, 2001) The experiment required both the demonstration and its socially organized witness.

Biology lagged behind physics in the experimental transformation of the seventeenth century for reasons Hall describes as conceptual rather than technical. Experimentation in biology was not deterred by technical difficulties alone, but by lack of imagination and the absence of conceptual frameworks that would give meaning to biological experiments. Santorio’s metabolic measurements, though they produced interesting information, carried little or no weight for or against the main strategic ideas of medicine and biology precisely because no framework yet existed to make them load-bearing.(Hall, A. Rupert, 1954)

The surgical tradition contributed its own instance of empiricism displacing authority before the century’s end. Hall identifies Ambroise Paré’s adoption of ligatures and dressed wounds in place of cauterization as a genuine revolt against received surgical doctrine, driven not by theory but by direct observation of outcomes: a technique sanctioned by centuries of authority gave way to bedside evidence.(Hall, A. Rupert, 1954)

The capacity of experiment to generate new knowledge depended on what questions investigators were asking. Francesco Redi’s 1670s experiments on spontaneous generation — demonstrating that decaying flesh generated maggots only when flies were allowed to settle on it — were decisive because they tested a specific and widely held theory against a simple manipulation.(Hall, A. Rupert, 1954) Marcello Malpighi’s microscopic observation of blood passing from arteries to veins through capillary vessels in a frog’s lung provided the final anatomical link that confirmed Harvey’s circulation — direct observation filling a gap that Harvey could only infer.(Hall, A. Rupert, 1954)

Bacon, Boyle, and the Royal Society: Experiment as a Social Practice

The seventeenth century produced not only experiments but a theory of how experiments should be conducted and reported. Dear’s analysis of this transformation identifies a fundamental shift in how experimental results were framed. The Aristotelian tradition expressed empirical facts as universal generalizations — “heavy bodies fall” — stating how things always behave rather than what happened on a particular occasion; the philosopher’s task was to explain such known truths, not to discover new ones.(Peter Dear, 2001) Bacon, Galileo, and the Royal Society each navigated this in different ways.

Francis Bacon reconceived natural philosophy as an enterprise productive of practical works rather than contemplative understanding, making technological benefit the criterion of genuine knowledge and declaring that “truth and usefulness are (in this kind) the very same things.”(Peter Dear, 2001) His critique of Aristotelian syllogistic logic held that genuine knowledge must begin with particular observations and ascend to universal axioms through induction — working from individual instances to general statements, not descending from pre-accepted universals to particulars.(Peter Dear, 2001) Rather than simply accumulating examples, his method required “rejections and exclusions” — eliminating all alternative explanations until only one remains.(Peter Dear, 2001)

Yet as Dear shows, Bacon’s experimental writing presented its results as universal recipes describing what always happens rather than what happened at a specific time and place, thereby bypassing the problem of getting readers to trust any particular claim. Bacon could describe “what happens” without tying it to a specific event requiring witnesses.(Peter Dear, 2001) Galileo similarly framed his inclined-plane experiments in the Discorsi as what happened “a full hundred times” — an appeal to common experience, not a report of specific historical events.(Peter Dear, 2001)

The Royal Society took a different path. Dear documents that when a Fellow of the Royal Society told his audience about an experiment, he typically told a story about an event that had happened in the past, to him, at a specific time and place.(Peter Dear, 2001) Robert Boyle’s experimental writing is characteristically circumstantial and particular, narrating events with precise detail about glassware sizes and exact procedures — a style designed to establish credibility through the weight of witnessed testimony.(Peter Dear, 2001) Shapin traces this specifically English experimental philosophy to the purposeful relocation of gentlemanly conversational codes into natural philosophy, constituting what he calls an “origins story” for modern science in which the social identity of witnesses and the conventions of gentlemanly discourse underwrote the credibility of experimental claims.(Shapin, 1994)

Thomas Hobbes argued that Boyle’s experimental program failed to achieve true natural philosophy because experiments could only display phenomena — they could not demonstrate necessary causes. Boyle could exhibit natural behaviors to which everyone might assent, but there was no way in experimental work to demonstrate what the causes of those behaviors must be: Hobbes could always supply several alternative interpretations, each as likely as Boyle’s.(Peter Dear, 2001) This dispute — whether experiment generates genuine explanations or only defensible descriptions — is not resolved by the history. It recurs in twentieth-century debates about what clinical trials prove.

Dear documents the central challenge that all seventeenth-century experimental science faced: converting specific, unusual experimental events into universally accepted truths. Trust in experimental narratives and universality of their outcomes had to be established simultaneously; when a phenomenon was produced only by unusual experimentation, the natural philosopher had to persuade readers both that it really happened and that it always happens.(Peter Dear, 2001)

The Reception of Harvey and the Spread of Experimental Norms

Experimental demonstration operated as a social process: Franciscus Sylvius arrived in Leiden in 1638 and within a year convinced Johannes Walaeus of Harvey’s circulation doctrine through private vivisection demonstrations witnessed by Bartholin, van Horn, and others.(French, 1994) Walaeus cut the tip from a dog’s heart ventricle, causing blood to spurt four feet and drenching observers.(French, 1994) The dramatic spurt made the point evident, as Walaeus remarked.(French, 1994)

John Primrose’s opposition to Harvey centered not on the logic of the argument but on the propriety of vivisection as a method. He argued that vivisection is unnatural because the violence of the procedure prevents it from showing the natural economy of the heart.(French, 1994) Walaeus dismissed this as illegitimate, accusing Primrose of holding “that the mind grasps things more surely than the eye can see them” — a phrase French identifies as currency among those who opposed experimental method and its sensory epistemology.(French, 1994) By mid-century, experiment had become so embedded in natural philosophy that purely verbal arguments were felt to be inadequate: Roger Drake upbraided Primrose for never having performed even the simplest ligature experiment.(French, 1994)

French’s central historiographic claim is that the important event in the Harvey affair was not that Harvey’s truth prevailed irresistibly, but that it came to be thought that truth — or the best available approximation to it — could be discerned by experiment.(French, 1994) The argument from experimental demonstration was not self-sufficient; it required social processes, consensus formation, and institutional authority to achieve persuasion. French argues explicitly that the truth of Harvey’s doctrines had no active power of itself — some other active thing was always needed to persuade people, and those mechanisms were the same ones that had produced consensus around falsities.(French, 1994) What changed in the mid-seventeenth century was that experiment became the mechanism people expected to be persuaded by.

French notes a more subtle epistemological point: vivisectional experiments could not claim Aristotelian demonstrative proof, which required linking the presence or absence of organs to causal relationships displayed across modes of life.(French, 1994) Yet Harvey’s use of vivisection helped form a consensus that experimental results stood apart from both rationalist systems and philosophical demonstrative truth.(French, 1994)

Claude Bernard and the Codification of Experimental Method

The systematic philosophical account of what medical experiment is and how it should be conducted came from Claude Bernard (1813–1878) in his Introduction to the Study of Experimental Medicine (1865). Bernard’s text was both a philosophical treatise on scientific method and a practical guide derived from decades of laboratory work, and it remains the clearest statement of what the experimental method in medicine requires.

Bernard defined experimental medicine as comprising three parts: physiology (knowledge of normal life), pathology (knowledge of disease causes), and therapeutics (use of medical agents to cure).(Bernard, 1927) He insisted that medicine must rest on physiology because pathological conditions can only be understood through knowledge of normal states.(Bernard, 1927) His most compact methodological formula was that reasoning is always the same whether in sciences studying living beings or inorganic bodies, but that the complexity of vital phenomena makes the principles of experimentation incomparably harder to apply to medicine than to physics.(Bernard, 1927) He set this claim within an explicit historical sequence: medicine had passed through empirical, systematic, nosological, and morphological stages before arriving at the experimental stage.(Bernard, 1927) Olmsted’s biographical account supplies a further institutional context: physiology as a separate science was practically nonexistent before 1850, and in Bernard’s early career the subject was still taught as an appendage to anatomy rather than an independent discipline.(Olmsted, 1938)

Bernard drew a distinction between observation (which shows) and experiment (which teaches), but argued the real difference is not between passive and active investigation but between reasoning with and without preconceived ideas. Both observation and experiment involve the activity of the mind; what distinguishes experiment is that a preconceived idea — a hypothesis — guides what one looks for.(Bernard, 1927) The experimenter who produced his working hypothesis first, and then observed without it, was the ideal: “the experiment is always devised with the help of a working hypothesis; the resulting observation is always made without preconceived idea.”(Bernard, 1927)

Henderson’s introduction to the 1927 English edition identifies the concept that gives Bernard’s method its specific claim: determinism. Bernard introduced determinism into physiology — the conviction that vital phenomena are governed by fixed laws, not by a capricious vital force.(Bernard, 1927) The concept of the milieu intérieur gave this determinism its physiological foundation: Bernard held that the constancy of the internal environment (the stability of blood, lymph, and interstitial fluid) is the very condition that makes free and independent life possible, allowing higher organisms to remain unaffected by the fluctuating external world.(Bernard, 1927) Vital phenomena, Bernard argued, possess rigorously determined physico-chemical conditions, but they subordinate themselves and succeed one another according to a law fixed in advance, an order that is irreducibly biological even while it obeys physico-chemical principles.(Bernard, 1927) Olmsted documents that the Introduction established scientific determinism as the foundation for experimental physiology: under identical conditions phenomena in living bodies are identical, as in inanimate bodies, directly opposing Cuvier, Bichat, and Magendie’s claims that a vital force nullified physico-chemical laws in living things.(Olmsted, 1938) This was the philosophical move that made medicine scientifically tractable in Bernard’s view: if biological phenomena are determined, they can be predicted, controlled, and experimentally manipulated.

Bernard’s account of his own discovery practice illustrates the method in action. Olmsted records that Magendie, near the end of his life, acknowledged Bernard as the natural heir to his chair: his last words to Bernard reportedly included the assurance that “my chair will come to you; with you I know it won’t fall.”(Olmsted, 1938) Bernard eventually surpassed his teacher’s reputation most dramatically with the glycogenic function of the liver, considered by Olmsted his greatest discovery (1848): Bernard demonstrated that the liver produced sugar, overturning the prevailing assumption that it only removed glucose from the blood and establishing that the body could manufacture the substance internally.(Olmsted, 1938)

Olmsted documents that Bernard’s pancreatic digestion work began with a chance observation that market rabbits, which had been fasting, had clear rather than cloudy urine. This led to a systematic investigation that culminated in demonstrating that the pancreas renders neutral fats absorbable — a major discovery arrived at through the disciplined pursuit of an unexpected result.(Olmsted, 1938) His first vasomotor experiment began similarly: he expected that cutting the sympathetic nerve would cool tissues by slowing blood combustion, but when he cut the nerve in a rabbit the head on the cut side grew warmer instead. Bernard’s response, as Olmsted records it, was characteristic: “I did as I always do; that is to say, I at once abandoned theories and hypotheses in order to observe and study the fact itself.”(Olmsted, 1938)

Olmsted, however, is also the source of a pointed counter-narrative. His analysis of Bernard’s early papers — on the chorda tympani, on gastric juice, on the spinal accessory nerve — finds a consistent failure to practice what Bernard later preached. In each case Bernard built theories on unverified assumptions rather than decisive experiments, “cluttering the literature with an impossible hypothesis of his own making.”(Olmsted, 1938) Magendie, whose extreme empiricism Bernard criticized — he compared himself to a scavenger picking up facts with no plan, proud of his complete lack of method(Olmsted, 1938) — was “a victim of this method or rather lack of method” in Bernard’s later assessment.(Olmsted, 1938) But Olmsted’s evidence suggests the distance between Magendie’s chaotic empiricism and Bernard’s principled experimentalism was shorter in practice than in theory.

Bernard’s philosophical position resisted easy categorization. Olmsted documents that he refused to affiliate with materialism, vitalism, or positivism as systems, dismissing all philosophical labels with the declaration that “systems do not exist in nature but only in men’s minds.”(Olmsted, 1938) He acknowledged vitalism’s kernel of truth — he agreed with vitalists that living beings exhibit phenomena peculiar to themselves and unknown in inorganic nature — while insisting those phenomena were experimentally determinable and not governed by an independent vital force.(Olmsted, 1938) He distinguished scientific determinism (the empirical reliability of biological phenomena for experimentation) from philosophical determinism (negation of human liberty), maintaining that the former is the necessary condition of the latter.(Olmsted, 1938)

The limits Bernard set on statistics are revealing.(Olmsted, 1938) He argued that “scientific law can be based only on certainty, on absolute determinism, not on probability,” and dismissed statistical methods as inadequate for genuine scientific law.(Olmsted, 1938) His concern was practical: statistical thinking might “bolster up medical empiricism and delay the development of a really scientific medicine.”(Olmsted, 1938)

Olmsted notes that Émile Zola in 1868 built his theory of naturalistic fiction directly on Bernard’s Introduction, attempting to apply the experimental method to creative writing in his Rougon-Macquart series.(Olmsted, 1938)

Canguilhem’s Critique: What the Experimental Method Cannot See

Georges Canguilhem (1904–1995) posed the deepest philosophical challenge to the premises of Bernard’s experimental medicine. Canguilhem’s target was not the value of experiment but the foundational claim, formalized by Auguste Comte building on Broussais, that disease is simply the excess or deficiency of excitation above or below a normal degree — that pathological phenomena differ from physiological ones only quantitatively, never qualitatively.(Canguilhem, 1966) Comte, Canguilhem documents, explicitly framed disease as a spontaneous experiment allowing comparison between abnormal and normal states, treating the pathological state as not producing truly new phenomena but only extending existing physiological phenomena in degree.(Canguilhem, 1966)

This is also the position that Bernard’s experimental method tacitly requires: if disease is just physiology pushed to extremes, then the laboratory can study pathological phenomena using the same deterministic methods it applies to normal ones. Canguilhem’s critique is that this is philosophically incoherent. His central thesis is that the pathological state does not consist in the absence of norms but is itself a norm of life — specifically, 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 in other conditions. This is a qualitative difference, not a quantitative one, and no amount of experimental measurement of physiological parameters can capture it.

Canguilhem also challenges the notion of “objective pathology” as logically incoherent. One can carry out research objectively — impartially, by experimental method — on a subject whose object cannot be conceived and constructed without being related to positive and negative qualifications, to value. The object of pathology is not a fact but a value.(Canguilhem, 1966) Canguilhem argues that medicine always exists because humans feel sick, not because doctors tell them of their illnesses: the sick person’s experience of being different is epistemologically prior to the physician’s objective knowledge, which historically derived from that experience.(Canguilhem, 1966) Foucault’s introduction to the English edition of Canguilhem’s work identifies this as the decisive move: Canguilhem opposed to phenomenology’s philosophy of meaning, subject, and the experienced thing a philosophy of error, concept, and the living being.(Canguilhem, 1966)

Canguilhem further notes that Bernard’s experimental method explicitly rejected statistics as inadequate for scientific law, insisting on absolute determinism. But even within experimental physiology, Canguilhem’s analysis suggests that what counts as “normal” is not a scientific fact determinable by experimental methods alone — it is a concept that cannot be reduced to an objective determination.(Canguilhem, 1966) Canguilhem does not deny that physiology can establish biological constants. He denies that those constants can, by themselves, define what is normal for a given living individual, because biological norms are individual rather than statistical: health is what permits an organism to meet the demands imposed on it, which may differ from population averages.(Canguilhem, 1966)

Foucault notes that for Canguilhem, discontinuity in the history of science is not a postulate or result but a method: science does not progress by the gradual discovery of an inscribed truth but through successive rectifications and corrections.(Canguilhem, 1966)

Contradictions and Open Questions

The history of the experimental method in medicine contains several contradictions that historians have not resolved.

The first concerns the relationship between experiment and theory. Bernard’s official position was that the experimenter should observe without preconceived ideas, but his own practice — as Olmsted demonstrates — regularly involved forming strong hypotheses and failing to test them with the rigor he advocated. His early papers made errors that stricter application of his own principles would have caught. Bernard’s formula was prescriptive rather than descriptive, and the question of how theory and observation actually interact in experimental practice remains contested.

The second concerns what experiment proves. Hobbes’s objection to Boyle — that experimental phenomena underdetermine their causes — was never fully answered. Harvey’s calculation that the blood circulates was devastating to Galenic hepatic blood production, but the mechanism of capillary circulation between arteries and veins, which Harvey inferred, was only directly observed by Malpighi after Harvey’s death. Bernard’s most famous discoveries regularly involved jumping from experimental results to theoretical conclusions that his measurements did not strictly support: Olmsted documents multiple cases where Bernard’s methods were insufficiently sensitive to detect what he claimed to have established or failed to detect. He was right about the theory and wrong about some of the measurements.

The third concerns statistics. Bernard’s categorical rejection of statistical reasoning as unsuited to genuine scientific law has been answered by twentieth-century medicine’s wholesale adoption of the randomized controlled trial — a statistical instrument that treats probability rather than determinism as the appropriate framework for medical knowledge. Whether the RCT captures what Bernard was after, or represents a fundamental change in what medicine considers to count as knowledge, is not a question the history of the experimental method resolves. The RCT was designed for situations where individual causal mechanisms cannot be identified, making it precisely the tool for the domain Bernard thought experiment should eventually eliminate.

The fourth contradiction is Sigerist’s. From the seventeenth century onward, he argues, medicine entered an inseparable alliance with the natural sciences that was fruitful when medicine retained its autonomy but dangerous when it reduced patients to mere natural objects and forgot its goal, which is cure.(Henry E. Sigerist, 1933) Bernard’s experimental medicine was built around the premise that medicine would eventually become scientific in the way physics is scientific — deterministic, predictive, exact. Whether that transformation would serve patients, or whether the demands of experimental science would distort the clinical encounter, Sigerist saw as permanently open.

Clinical and Phenomenological Critiques of Experimental Medicine

Bernard’s hope that medicine would eventually achieve the predictive exactitude of physics became the working premise of twentieth-century scientific medicine. The shift was visible even in the late nineteenth century. Thomas Huxley’s 1876 address at the opening of Johns Hopkins, the institution that would do more than any other to standardize research-based medical education in America, endorsed professional competence as the criterion of right medical conduct, displacing earlier appeals to moral character or classical learning with a standard grounded in scientific training and demonstrated skill.(Jonsen, 2000) Eric Cassell (b. 1928) traces the institutional moment at which this premise became orthodoxy: it was not with the Flexner Report of 1910, which gave medical science a central place in physician education, but only after the Second World War that experimental medical science done by a scientifically oriented faculty became the dominant force in American medical schools, displacing clinicians and primary concern for patients.(Cassell, 2014) The philosophical framework that accompanied this shift was positivism — the claim that only scientifically verifiable facts about disease and human biology should guide diagnosis and treatment, a move whose internal logic implied that the neophyte physician, armed with the same verified knowledge, would be as effective as the experienced attending.(Cassell, 2014) Bernard’s determinism had become an institutional program.

Cassell’s central argument is that this program is epistemologically incoherent as a description of what clinical medicine actually is. The word “judgment” was replaced in post-Feinstein clinical epidemiology by the words “decision-making,” which allowed quantitative rules — sensitivity, specificity, positive predictive value, likelihood ratios — to substitute for the necessarily qualitative reasoning required in most clinical situations.(Cassell, 2014) But judgment, Cassell insists, is by definition an opinion arrived at by applying general information to a specific situation, and there can therefore be no rules or quantitative methods for arriving at one.(Cassell, 2014) The call for “evidence-based medicine” and the proliferation of clinical guidelines are recurring attempts to get around the necessity for the judgment of individual doctors — attempts that are, in Cassell’s view, impossible given the inherently public and social character of that judgment.(Cassell, 2014)

The statistical instrument that most completely embodies the aspiration Bernard distrusted — the randomized controlled trial — runs up against a specific epistemic obstacle that Cassell identifies in geometric terms. Statistical methods eliminate individual differences to arrive at general truths, but the individual patient in front of the clinician is precisely the point that lies off the regression line.(Cassell, 2014) Florence Nightingale had made essentially the same point in 1860: “a want of the habit of observing conditions and an inveterate habit of taking averages are each of them often equally misleading.”(Cassell, 2014) Clinical epidemiology as a discipline never came to terms with the ineliminable subjectivity of information from patients and physicians, nor with the difficulties posed by the intricacy and variability of information unfolding over time.(Cassell, 2014) Clinical medicine draws heavily on science and its methods, but the knowledge and ability to think clinically cannot be duplicated or rendered irrelevant by scientific knowledge of the same things, because that is a different perspective — one that is unique in being knowledge through time.(Cassell, 2014)

Kathryn Montgomery makes the same point in the language of Aristotelian philosophy. Medicine is not itself a science but a practice that uses science: it is characterized by the clinical judgment required to apply general rules to particular patients.(Montgomery, 2006) The misrepresentation of medicine as a Newtonian or Galilean science — as a body of invariant, objective, and always replicable knowledge — has had damaging practical consequences: a harsh and often brutal medical education, unnecessarily impersonal clinical practice, dissatisfied patients, and disheartened physicians.(Montgomery, 2006) Montgomery names this collective misunderstanding an “epistemological scotoma” — a blindness of which the knower is unaware.(Montgomery, 2006) Like history or evolutionary biology, she argues, clinical medicine is fated to be a retrospective, narrative investigation and not a Newtonian science.(Montgomery, 2006) The statistical knowledge that the RCT provides fails to resolve this problem: no one survives 82 percent — survivors survive entirely, those who die are completely dead — and statistics say nothing about any one particular person.(Montgomery, 2006) The cultural dimension of the problem is visible in breast cancer treatment: French surgeons were performing lumpectomies long before American ones, because “the French like breasts” while the Americans “like randomized clinical trials.”(Montgomery, 2006) What counts as sufficient experimental evidence is not determined by the logic of the experiment alone.

The phenomenological tradition offers a deeper diagnosis. Hans-Georg Gadamer argues that modern science and its ideal of objectification demands of doctors, patients, and responsible citizens alike a violent estrangement from themselves — from their own lived bodily experience.(Gadamer, 1996) The clinical investigation, which reduces the individual patient to a multiplicity of data that could be collated on a card index, raises the question of whether the unique value of the individual is properly recognized in this process.(Gadamer, 1996) Even the standard values — the laboratory reference ranges and population norms — that structure clinical interpretation are, in Gadamer’s assessment, one of the principal sources of error in established medicine, because they replace the direct observation of and listening to the patient with numbers read off technologically sophisticated measuring instruments.(Gadamer, 1996) The theory-practice gap that Bernard believed the experimental method would eventually close is, for Gadamer, constitutive of medicine rather than a defect to be remedied: once science has provided doctors with general laws, causal mechanisms, and principles, they must still discover what is the right thing to do in each particular case, and this is something that hardly seems predictable or knowable in advance.(Gadamer, 1996) Edmund Husserl’s concept of the life-world provides the alternative orientation: all methodological science encounters limits with respect to what it can hope to achieve, and doctor and patient are partners in a shared world that cannot be reduced to resistance and mastery.(Gadamer, 1996)

Havi Carel (b. 1970) extends this line of argument to a precise clinical claim. The experimental method in medicine, as practiced in clinical trials, takes disease — physiological dysfunction — as its object of investigation. But the experience of illness (how the disease is experienced by the patient) is a distinct phenomenon that cannot be captured by the same tools: it is not enough to see illness as an entity in the world that can be studied with the tools of science; illness must also be studied as a lived experience, whose existential, ethical, and social dimensions fall outside the scope of controlled experiment.(Carel, 2016) The gap between objective disease markers and patients’ lived experience is not merely philosophically interesting but clinically consequential: interventions designed to restore objectively measured function — an increase in a lung function value, a reduction in hospital admissions — do not necessarily correlate with a perceived increase in health or well-being by the patient.(Carel, 2016) The “disability paradox” documents this gap empirically: patients with severe disease or disability, including renal patients tethered to dialysis machines three times a week, report levels of well-being comparable to healthy controls.(Carel, 2016) Modern medicine’s commitment to the body as a biological object, Carel argues, has generated enormous scientific progress but also a systematic neglect of illness experience, visible in hospital design, communication problems between healthcare staff and patients, and low patient satisfaction.(Carel, 2016) Patients are not merely the subjects of experimental medicine’s findings; they are knowers of their own experience, and the dismissal of that knowledge constitutes what Carel and Ian James Kidd call testimonial injustice — a wrong done to the ill person in their capacity as knower, effected through the presumptive attribution of cognitive unreliability that downgrades the credibility of ill people’s testimonies.(Carel, 2016)

The convergence across these four thinkers — Cassell from clinical medicine, Montgomery from philosophy of medicine, Gadamer from hermeneutics, Carel from phenomenology — is striking. Each identifies the same structural problem: the experimental method, as codified in the RCT and the EBM apparatus built around it, treats the population as the unit of valid knowledge and statistical probability as the appropriate form of medical truth. But medicine as actually practiced addresses particular patients in their irreducible individuality, and requires the kind of practical wisdom that Aristotle called phronesis — the flexible, interpretive capacity to determine the best action when knowledge depends on circumstance.(Montgomery, 2006) Bernard’s dismissal of statistics as “bolstering medical empiricism” has been inverted by twentieth-century medicine, which treats statistical probability not as a deficiency to be transcended but as the highest available form of medical knowledge. Whether the experimental method’s population-level findings can be translated to individual clinical practice without an irreducible act of interpretation that the method itself cannot supply remains an unresolved question in the philosophy of medicine.

See Also

• anatomy
• vivisection
• William Harvey
• Claude Bernard
• galen
• humoral-theory
• physiology
• clinical-observation
• randomized-controlled-trial
• vitalism
• determinism-in-medicine
• normal-and-pathological
• bacteriology

Editorial Notes

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

Contradictions and Open Questions