Avicenna’s Seven Rules for Testing Drugs

Somewhere in the second book of The Canon of Medicine, the great medical encyclopedia that the Persian physician Avicenna (Ibn Sīnā, about 980–1037) finished at Hamadan in 1025, sits a short chapter on how a physician can find out, by experiment, how strong a medicine really is. It sets out seven conditions. A drug has to be pure; it has to be tried on a single, uncomplicated illness; it has to be tried against contrary conditions; its strength has to be matched to the disease, starting weak; the time it takes to act has to be watched; its effect has to repeat in all or most cases; and it has to be tested on human beings, because an animal may react differently. Read today, the list sounds strikingly familiar, and modern writers have called it everything from “11th century rules for assessing the effects of drugs” to “the earliest known treatise related to clinical trials.”

This page walks through that chapter rule by rule, explaining each in plain language and setting it beside the modern idea it most resembles: the standardised drug, the inclusion criteria of a study, dose escalation, the time course of a drug’s action, reproducibility, and the gap between animal and human results. It also looks at what Avicenna inherited from the Greek physician Galen, at what the rules leave out — no control group, no randomisation, no blinding — and at how historians and physicians writing since 2009 have read them, each view attributed to the writers who hold it. The rules are quoted only in translation, following the English of Nasser, Tibi and Savage-Smith (2009); elsewhere they are paraphrased.

Table of Contents

  1. Knowing a Drug by Experiment
  2. Reason, Experience and the Physician
  3. Rules One and Two: A Pure Drug, a Single Disease
  4. Rule Three: Contrary Conditions
  5. Rule Four: Starting Weak and Increasing
  6. Rules Five and Six: Timing and Consistency
  7. Rule Seven: Humans, Not Animals
  8. What Came from Galen
  9. What the Rules Did Not Include
  10. Modern Readings of the Rules
  11. Key Research Papers
  12. Connections
  13. Featured Videos

1. Knowing a Drug by Experiment

The Canon is divided into five books. The first sets out the general principles of medicine, the third and fourth describe diseases, and the fifth is a formulary of some 650 compound remedies. The second book is the materia medica: a catalogue of roughly 800 “simple” drugs — single substances, most of them plants, with many animal and mineral products among them. According to the summary of its structure given by the historian Danielle Jacquart and others, the book opens with a general part on the rules for drugs and the “powers” of medicines, and only then turns to the alphabetical list, in which each entry records the drug’s name, how to recognise a good specimen, its primary qualities (hot or cold, wet or dry), its general actions, its uses for different kinds of disease, and the substances that can stand in for it.

The drug-testing chapter belongs to that opening general part. Its title, in the translation used by Nasser, Tibi and Savage-Smith, is “On knowledge of the potency of drugs through experimentation.” Avicenna begins by saying that experimenting leads to certain knowledge of a medicine’s potency — but only “after taking into consideration certain conditions.” The seven rules are those conditions. They are not a description of any study he ran; they are a set of requirements that an experiment on a drug has to meet before its result can be trusted.

That framing matters for everything that follows. Avicenna was not asking whether a remedy “works” in the modern, statistical sense. He was asking how a physician can be sure that a change seen in a patient was produced by the drug itself, through its own nature, and not by something else: an impurity, a second illness, the season, chance, or the patient’s own recovery. Seen that way, the seven conditions are a list of the ways a drug test can mislead, each paired with a way of closing the gap.

Why a medieval physician needed rules at all

A physician of Avicenna’s time inherited hundreds of drug descriptions from Greek, Persian, Indian and Arabic authorities, and the descriptions often disagreed. Book 5 of the Canon shows him comparing versions of the same recipe and judging between them, and according to the summary of his formulary he favoured remedies proven by experience and warned that compound drugs could act unexpectedly, or more strongly, than their ingredients. The testing chapter is the theoretical side of the same habit: a method for checking an inherited claim against what actually happens in patients.

Back to Table of Contents

2. Reason, Experience and the Physician

Medieval Arabic medicine recognised two ways of knowing what a drug does. The first was reasoning: working out a drug’s powers from its known qualities and from analogy with other substances, inside the theory of elements, humours and temperaments that the Canon takes over from Galen and fits into Aristotle’s natural philosophy. In that theory every drug has a degree of heat or cold, dryness or moisture, and it acts by pushing the body’s balance of qualities one way or the other. The second way was experience — trying the drug and watching. The testing chapter is Avicenna’s account of how to make the second way reliable.

The two were not rivals for him. The rules themselves speak the language of the theory: a drug can carry an “acquired” heat or cold, a disease has a “strength,” and a drug’s heat can be weaker than the cold of the disease it is meant to correct. Experience was the check on reasoning, and reasoning told the physician what to look for in the experience. The historian Jacquart describes Arabic pharmacology between the ninth and thirteenth centuries as building on Galen and Dioscorides and moving between two approaches — grading drugs by degrees of their primary qualities, and recognising “specific” powers that belong to a substance’s whole form and cannot be predicted from its qualities. Powers of the second kind could only be known by trial.

The physician as observer

Ghaffari and colleagues (2022) gathered passages across the Canon in which Avicenna appeals to clinical experience and observation of patients, arguing that the book is not only a work of inherited theory. Shoja and colleagues (2011) look further back, at the logic behind the method: they trace influences from the Islamic legal idea of consensus (ijmāʿ) and from Stoic logic, and suggest that Avicennian logic may be an indirect ancestor of evidence-based medicine. Both are modern interpretations, and both treat the testing chapter as the place where Avicenna’s philosophy and his medicine meet most directly.

Back to Table of Contents

3. Rules One and Two: A Pure Drug, a Single Disease

Rule one: the drug free of any acquired quality

The first condition, in translation, is that the drug be free from any acquired quality. Avicenna had in mind changes a substance picks up after it is gathered: being heated or chilled, or being contaminated by being stored next to something else. In the humoral framework such a change could alter the drug’s apparent heat or cold, so a test on a spoiled or altered sample would report the wrong nature for the drug itself.

The modern counterpart is the standardised test substance. A present-day drug trial is run on a product of known identity, purity and strength, made and stored under controlled conditions, so that every patient receives the same thing and a result can be traced back to one compound. Herbal research meets the same problem in a sharper form: plant material varies with species, growing conditions, harvest, drying and storage, and studies of the same herb can disagree partly because they did not test the same material. Avicenna’s first rule names the problem in the vocabulary of his time.

Rule two: a single, not a composite, disease

The second condition is that the drug be tried on a single, not a composite, disease. If a patient has two illnesses at once, an improvement cannot be credited to the drug’s action on either of them with any confidence: the change may belong to the other illness, or to the way the two interact.

The modern counterpart is the set of inclusion and exclusion criteria that define who enters a study. Trials usually begin with a narrowly defined group of patients who share one diagnosis, precisely so that the effect being measured is not blurred by other conditions; broader groups come later, once the basic effect is known. The rule also mirrors the Canon’s own structure. Book 2 classifies each simple drug’s specific uses against a grid of disease types, and a drug can only be placed on that grid if its effect on each kind of disease has been seen separately.

Back to Table of Contents

4. Rule Three: Contrary Conditions

The third condition asks the physician to test the drug on two contrary conditions. It is the hardest of the seven to read, and translators phrase it differently. The aim, as Nasser, Tibi and Savage-Smith explain it, is to separate what a drug does by its own nature from what it does only incidentally. A drug may relieve one illness because its essential quality opposes that illness, and relieve a quite different illness only by accident — for instance by acting on a symptom the two share, or by removing some secondary cause. Watching the drug against conditions of opposite kinds helps the physician decide which of its effects truly belong to it.

Avicenna’s distinction between “essential” and “accidental” comes from Aristotle’s logic, and it runs through the whole chapter: rules five and six return to it. In plain terms, the question is whether the drug is treating the disease or only its signs.

The modern echo

Modern pharmacology asks the same question in different words. A medicine that lowers a fever, eases a pain or quiets a cough may change how a patient feels without changing the course of the underlying illness, and trials are designed to tell symptom relief apart from an effect on the disease itself. Researchers also distinguish a drug’s main action from its side actions, which can show up as benefits in one condition and harms in another. Avicenna did not have the concepts of receptors, mechanisms or end points, but the third rule is an early attempt to stop a physician from mistaking a drug’s incidental effect for its real one.

Back to Table of Contents

5. Rule Four: Starting Weak and Increasing

The fourth condition joins two ideas. First, the strength of the drug has to correspond to the strength of the disease: in the humoral picture, a drug whose heat is weaker than the cold of the illness will show no effect at all, and a physician who concludes from this that the drug is useless has made a mistake. Second, the test begins with the weakest dose and increases it gradually, until the drug’s power is known.

The first idea is a warning against false negatives — failing to see an effect because too little of the drug was given. The second is a safeguard against harm and a way of finding where the effect begins. Avicenna knew that many of his drugs were dangerous. His own entry on opium in Book 2, as described by Heydari, Hashempur and Zargaran (2013), lists its pain-relieving, sleep-inducing and cough-suppressing effects alongside its depression of breathing, and calls it a potential poison. A rule that starts low and climbs slowly fits a materia medica in which the line between remedy and poison was thin.

Dose escalation then and now

The modern counterpart is dose escalation, the basic design of the first studies of a new medicine in people. Volunteers or patients receive a low dose; if it is tolerated, the next group receives more, and the process continues until the dose range that produces an effect, and the dose at which side effects appear, are both known. Alongside it sits the dose–response relationship, the idea that the size of a drug’s effect depends on how much is given. Paracelsus, five centuries later, would be remembered for the saying that the dose makes the poison; Avicenna’s fourth rule shows the same concern written into a testing procedure. The history of dosing and the ancient and medieval drugs Avicenna catalogued is told on this wing’s page on opium, saffron and the long life of the Canon; no historical quantities are given on either page.

Back to Table of Contents

6. Rules Five and Six: Timing and Consistency

Rule five: the time to effect

The fifth condition is to observe how long the drug takes to act. In Avicenna’s reasoning, an effect that follows immediately shows that the drug acted on the disease itself; an effect that is delayed, or that at first runs contrary to what comes later, leaves the physician uncertain, because it may be accidental rather than the drug’s own action. The rule again aims at the difference between an essential and an accidental effect, this time using time as the test.

The modern counterpart is the time course of a drug’s action — how quickly it is absorbed, how soon its effect begins, when the effect peaks and how long it lasts — the field known today as pharmacokinetics and pharmacodynamics. Modern pharmacology would not accept Avicenna’s conclusion as a general law, since many medicines act slowly by their nature. But the underlying point, that the timing of a change is evidence about its cause, and that improvement long after a dose may simply be recovery, remains part of how drug effects are judged.

Rule six: the same result in all or most cases

The sixth condition is that the drug’s effect be the same in all, or at least most, of the cases in which it is tried. If it is not, in the translation used by Nasser and colleagues, the effect is accidental. One good result proves nothing; a result that repeats is evidence.

The modern counterpart is reproducibility, and behind it the whole apparatus of counting. Today a single dramatic recovery is treated as an anecdote, a series of cases as a signal, and a result that holds up across many patients and across independent studies as stronger evidence. Avicenna had no statistics and no way of saying how many cases were enough, but “all or most” expresses the same intuition that a drug’s true effect shows itself as a pattern across patients rather than in any one of them.

Back to Table of Contents

7. Rule Seven: Humans, Not Animals

The last condition is that the experiment be made on the human body. A drug tried on an animal, Avicenna reasoned, might act differently there than in a person, so a test on an animal could not settle what the drug does in a patient. In the humoral framework the reason was plain: animals have different temperaments from human beings, so the same drug could be relatively hotter or colder in one than in the other.

This is the rule modern commentators single out most often. Nasser, Tibi and Savage-Smith write that the seventh rule “remains very relevant,” given how often the results of animal studies fail to match what is later found in people. Modern medicine does use animal studies, mainly to look for toxicity and to explore mechanisms before a drug reaches human volunteers. But it treats them as a preliminary step, and a medicine is only judged on what it does in human beings. Avicenna’s seventh rule makes the second half of that point a thousand years earlier, without the first.

A rule written for the patient

The seventh rule also says something about where medical knowledge came from in Avicenna’s world. A court physician like him — he served rulers from Bukhara to Isfahan, as told on the Life and Times page — learned about drugs mainly by giving them to patients and watching what happened. The testing chapter turns that everyday practice into a method. It says nothing about consent or the protection of the people a drug is tried on; those ideas belong to the twentieth century.

Back to Table of Contents

8. What Came from Galen

Avicenna did not invent drug testing from nothing. The Canon is, in the words its historians use, a Galenic medical summa: it takes the medicine of Galen of Pergamon (second century), as developed by later Greek and Arabic physicians, and orders it within Aristotle’s philosophy. Sadeghi and colleagues (2020) count about 300 citations of Galen in the Canon. They show Avicenna treating him with respect, confirming much of his teaching, and openly rejecting some of his pharmaceutical opinions and some of his physiology.

The testing rules carry the same mixed inheritance. The historian of science A. C. Crombie, whose 1971 study is followed by much later writing, judged that the rules were taken partly from Galen. Galen had written at length on simple drugs and on how their powers could be known, and the language of qualities, degrees and essential and accidental effects in Avicenna’s chapter is the shared language of that Greek tradition. Behind both stands Dioscorides, whose first-century De Materia Medica gave Greek and Arabic pharmacology its catalogue of plants; Jacquart names Galen and Dioscorides together as the foundations Arabic pharmacology built on.

What was Avicenna’s own

Where the line falls between Galen’s contribution and Avicenna’s is a question for specialists who can compare the Greek and Arabic texts, and the sources used for this page do not settle it rule by rule. What the modern commentators credit to Avicenna is the form: a short, numbered set of conditions, gathered in one chapter, stated as requirements for any experiment on a drug, and placed at the head of the most widely taught medical textbook of the following centuries. Through the Latin Canon, translated at Toledo in the twelfth century and taught in European universities for hundreds of years, that compact form reached readers who never saw Galen’s longer treatments.

Back to Table of Contents

9. What the Rules Did Not Include

Reading the rules beside a modern trial protocol is useful precisely because the differences are so clear. Four of them stand out.

A fifth limitation runs underneath the other four. The rules measure a drug against the theory of hot, cold, wet and dry, so the “strength” of a drug and of a disease are humoral quantities, not doses and blood levels. A test could be carried out perfectly by Avicenna’s standards and still be interpreted through a theory of the body that modern physiology has abandoned.

Never a trial, as the sources agree

None of the sources examined for this page says that Avicenna carried out a clinical trial, and none of the modern papers claims his rules describe a controlled one. Nasser, Tibi and Savage-Smith present them as rules for assessing the effects of drugs. Even Sajadi and colleagues, who use the phrase “clinical trial” in their title, describe the chapter as a treatise related to clinical trials. The rules are conditions for an experiment, not the record of one.

Back to Table of Contents

10. Modern Readings of the Rules

Sajadi, Mansouri and Sajadi (2009)

Writing in the Annals of Internal Medicine, Mohammad Sajadi, Davood Mansouri and Mohamad-Reza Sajadi described the seven conditions for “the recognition of the strengths of the characteristics of medicines through experimentation” and provided their own translation of the chapter. In their reading it is “the earliest known treatise related to clinical trials.” That is the authors’ judgement, and the qualifier in it — “related to” — is part of the claim. The paper drew a published comment from M. H. Azizi in the Archives of Iranian Medicine later the same year.

Nasser, Tibi and Savage-Smith (2009)

In the Journal of the Royal Society of Medicine, Mona Nasser, Aida Tibi and the historian of Islamic medicine Emilie Savage-Smith translated the chapter and commented on how several of the rules relate to present-day thinking about the assessment of treatments. Their emphasis falls on the seventh rule, which they call still very relevant in light of the frequent disagreement between animal and human studies. They read the chapter as a set of eleventh-century rules for assessing the effects of drugs, not as a trial.

Shoja and colleagues (2011)

Mohammadali Shoja and co-authors, writing in the International Journal of Cardiology, set the rules within Avicenna’s logic. They trace the influence of the legal concept of consensus and of Stoic logic on his thinking and argue that Avicennian logic may be an indirect source of the reasoning behind evidence-based medicine. The argument is about a line of ideas, offered by the authors as a possibility.

Other readings

Jamal Moosavi (2009) describes medical teaching from the eleventh to the seventeenth century as turning on the pivot of Avicenna’s medicine, and Zargaran and colleagues’ 2012 biographical note in the Journal of Neurology is one of several short tributes to him in clinical journals. The Arabic commentators of the twelfth to fourteenth centuries also read the Canon critically: Nahyan Fancy (2020) shows Fakhr al-Dīn al-Rāzī and Ibn al-Nafīs questioning and testing its claims rather than simply repeating them.

The long road to the controlled trial

Between Avicenna’s chapter and the modern randomised trial lie many centuries. One landmark often told is the 1747 comparison of treatments for scurvy by the Scottish naval surgeon James Lind, described on the site’s Vitamin C history page; the control groups, randomisation and blinding missing from Avicenna’s list became standard only in the twentieth century. Avicenna’s rules sit at the start of that road, as a statement of what any trustworthy test of a medicine has to rule out. The story of how the Canon itself travelled into the universities of Europe is told on this wing’s Canon of Medicine page and its legacy page.

Back to Table of Contents

Key Research Papers

  1. Nasser M, Tibi A, Savage-Smith E. Ibn Sina’s Canon of Medicine: 11th century rules for assessing the effects of drugs. J R Soc Med. 2009;102(2):78-80. PubMed PMID: 19208873
  2. Sajadi MM, Mansouri D, Sajadi MR. Ibn Sina and the clinical trial. Ann Intern Med. 2009;150(9):640-3. PubMed PMID: 19414844
  3. Azizi MH. “Ibn Sina and the clinical trial”. Arch Iran Med. 2009;12(4):435-6. PubMed PMID: 19566367 (comment on Sajadi et al. 2009)
  4. Shoja MM, Rashidi MR, Tubbs RS, Etemadi J, Abbasnejad F, Agutter PS. Legacy of Avicenna and evidence-based medicine. Int J Cardiol. 2011;150(3):243-6. PubMed PMID: 21093081
  5. Ghaffari F, Taheri M, Meyari A, Karimi Y, Naseri M. Avicenna and clinical experiences in Canon of Medicine. J Med Life. 2022;15(2):168-173. PubMed PMID: 35419109
  6. Sadeghi S, Ghaffari F, Heydarirad G, Alizadeh M. Galen’s place in Avicenna’s The Canon of Medicine: Respect, confirmation and criticism. J Integr Med. 2020;18(1):21-25. PubMed PMID: 31787564
  7. Heydari M, Hashempur MH, Zargaran A. Medicinal aspects of opium as described in Avicenna’s Canon of Medicine. Acta Med Hist Adriat. 2013;11(1):101-12. PubMed PMID: 23883087
  8. Moosavi J. The place of avicenna in the history of medicine. Avicenna J Med Biotechnol. 2009;1(1):3-8. PubMed PMID: 23407771
  9. Zargaran A, Mehdizadeh A, Zarshenas MM, Mohagheghzadeh A. Avicenna (980-1037 AD). J Neurol. 2012;259(2):389-90. PubMed PMID: 21887514
  10. Fancy N. Verification and Utility in the Arabic Commentaries on the Canon of Medicine: Examples from the Works of Fakhr al-Dīn al-Rāzī (d. 1210) and Ibn al-Nafīs (d. 1288). J Hist Med Allied Sci. 2020;75(4):361-382. PubMed PMID: 32974673
  11. Jacquart D. Islamic Pharmacology in the Middle Ages: Theories and Substances. European Review. 2008;16(2):219-227. DOI: 10.1017/S1062798708000215

PubMed Topic Searches

  1. https://pubmed.ncbi.nlm.nih.gov/?term=Ibn+Sina+drug+testing
  2. https://pubmed.ncbi.nlm.nih.gov/?term=Ibn+Sina+clinical+trial
  3. https://pubmed.ncbi.nlm.nih.gov/?term=Avicenna+Canon+of+Medicine
  4. https://pubmed.ncbi.nlm.nih.gov/?term=Avicenna+evidence-based+medicine

Further Reading

Back to Table of Contents

Connections

Back to Table of Contents