This is a statement of my views about homeopathy, with a focus on the missing links in the evidential chain.


The scientific consensus is that there is no reason to suppose homeopathy should work, no way it can work, and no proof it does work.

Homeopaths claim that science dismisses homeopathy because it cannot explain it. This is not true.

Science has a complete and coherent explanation for all the observations offered, which is consistent both internally and externally (with other scientific knowledge). The scientific explanation does not require any ad hoc hypotheses or speculative mechanisms. Every part of the scientific explanation is consistent with robustly empirically established phenomena.

Homeopaths, by contrast, rely on explanations which are incomplete, incoherent, internally self-contradictory, inconsistent with other branches of knowledge, and rely heavily on ad hoc hypotheses, speculative rationales and in some cases refuted concepts.

The scientific explanation for the observed facts includes:

  • Nonspecific effects, including
    • Placebo effects
    • Expectation effects
    • Regression toward the mean
    • Natural history of disease
  • Biases including
    • Observer bias
    • Experimenter bias
    • Publication bias

Collectively, this forms the null hypothesis in any trial of homeopathy. No trial has convincingly refuted the null hypothesis.

All these are well known factors which complicate all attempts to assess any form of treatment. There is no credible reason to suppose that homeopathy, alone, would be immune to these.

For example, you might have arthritis which causes pain; the pain varies over time. You are likely to consult a quack if the pain is particularly bad – and that is, according to the cyclical nature of the condition, exactly when you’d expect it to start to diminish. Quacks have a number of escape hatches they use when this does not happen: they might claim a “healing crisis” for example, or change the “remedy”. This will continue until the inevitable happens and the pain subsides of its own accord- in which case both the quack and you will tend to attribute the improvement to homeopathy. This is why “patient experience” is a notoriously unreliable guide, and it’s also why homeopaths seek to reverse the hierarchy of evidence to put individual patient experience at the top.


Homeopathy is a field that was made up out of whole cloth by Samuel Hahnemann in the late 18th and early 19th Centuries. At the time, conventional medicine was probably more likely to kill you than cure you. Bloodletting was a common and fashionable treatment. This was the time of heroic medicine and Hahnemann grew increasingly uneasy of it. His homeopathy was proposed as a gentler system – which it was. It is clear from the documentation that Hahnemann was sincere and although his ideas seemed odd they were no less odd than many contemporary medical practices. It is worth noting that the substance which insipired the idea was cinchona bark as a cure for malaria; Hahnemann found that consuming the bark caused symptoms like those of malaria, and that in dilute form the essence of the bark was in use to give effective relief for the symptoms. We now know that the active ingredient in cinchona bark is quinine.

From this and other work, Hahnemann proposed a “law of similars” (similia similibus curentur, let like cure like) which ipse diit axiom is a founding principle of homeopathy. He also proposed the idea that serial dilution made the remedy stronger, and that this serial dilution must be accompanied by a special form of shaking known as succussion. Finally, Hahnemann proposed that diseass were caused by miasms, disruptions of the body’s vital force.

What we did not know then but know now

Here’s a partial list of relevant scientific concepts that were at that time either unknown, poorly understood or not yet generally accepted:

In addition vaccination was in its infancy and the mechanism by which it worked was a mystery; superficially, it had some similarities with homeopathy though now of course we understand the mechanism of vaccination and know that it is not similar at all.

What we now know

Indicative diagram showing the intersection between medicine and homeopathy. Hahnemann and modern medicine more or less touch at the work of Edward Jenner, pioneer of inoculation.

Since Hahnemann’s time we have discovered a lot about the nature of the world. We know that infectious diseases are caused by pathogens, that the smallest amount of a substance is one molecule. We understand more than ever before about the chemistry of the human body. But of course we also know that there is a lot still to learn, and the scientific method has been developed over the centuries as a reliable way of refining our understanding of the world around us. The wonderful Richard Feynmann describes the scientific method well: the development of ever more complex and detailed explanations of the world, with occasional giant leaps that simplify everything again. Like peeling an onion, we expose successively profound understanding of how matter and life is governed.

The hallmark of the scientific method is that when something is disproven, it is discarded without mercy. Wikipedia’s a list ofsuperseded scientific theories includes a decent number of medical hypotheses including the “miasma theory of disease”. Originally it was thought that atoms were indivisible; now we know that is not so.

Homeopathy lacks this constant process of reappraisal. Homeopathy now would be substantially recoggnisable to Hahnemann, though he might well dispute some of its lunatic fringes.

The law of similars

One of the things that all this research and refinement of theories has not done is to validate the law of similars.

Given that Hahnemann, in common with everyone else of his day, knew next to nothing of most of the underlying causes of disease, it is necessary to interpret the law of similars and to decide whether it is proposed as a relief of the underlying cause, or of the symptoms.

If homeopathy is proposed as a relief for the underlying cause of disease, then it falls flat immediately. Diseases have so many disparate causes that it is simply not credible that any meaningful number will be cured by a dilution of something that causes superficially similar symptoms. Several diseases can exhibit similar symptoms. It can be difficult in some cases to distinguish, especially in the early stages, even the simplest difference, between a viral and a bacterial infection.

Oscillococcinum is an instructive example to consider. Oscillococcinum is a homeopathic preparation of the oscillicoccus bacterium which caused Spanish flu, as observed and diagnosed by the French physician Joseph Roy; he then went on to hypothesis this bacterium as the cause of numerous diseases, and identified the liver of a duck as being a good source for preparation of the mother tincture.

We now know that influenza is caused by a virus, not a bacterium, and that the oscillococcum bacterium does not exist. Roy stated that the cause was the same for such disorders as flu, herpes and shingles. This is not true, these are due separate and distinct viruses. If oscillococcinum were a conventional medical product then it would not be licensed because the theoretical basis is simply wrong. But we’re dealing with homeopathy here and it has more in common with literal biblical creationism than with science. Oscillococcinum is in the books, therefore it is still produced and in fact is said to be the biggest selling homeopathic remedy on the planet. If you have flu symptoms, a homeopath will, like as not, recommend that you take this remedy prepared from the liver of a duck which purportedly contains the (non-existent) bacterium which we now know doesn’t cause flu.

The law of infinitesimals

Water memory: Nature's take.

Hahnemann knew nothing of the existence of atoms. Modern homeopaths have no such excuse.

If you take 18g of water it contains around 6×1023 molecules of water. From this we can see that a 30C dilution, that is, serial dilution by one part in a hundred, performed thirty times, one of Hahnemann’s recommended dilutions and still in common use in homeopathy, results in a statistically improbable one in ten to the thirty seventh power chance of finding a single molecule of the original substance. Higher dilutions are in use. Oscillococcinum, one of the most popular remedies on sale, is usually sold at 200C dilutions, or one part in 10400. The known universe is estimated to contain somewhere of the order of 1079 atoms. Imagine taking one atom and diluting it in the entire universe, then taking one atom of the result and diluting that in an entire universe, then doing the same thing another two or three times, and you’re getting close. To any scientist, these levels of dilution are absurd.

Homeopaths have advanced post-facto rationalisations based on the idea that water has memory. Scientists are sceptical of this because it implies that you can re-create the effect of a molecule from its “energy imprint” on something many orders of magnitude smaller. A preparation based on a nosode, a disease organism, will have to reproduce the effect of the organism. Human DNA has around 200 billion atoms, bacterial DNA many fewer of course but a bacterium is made of more than a single strand of DNA. We know from the field of signal processing that it is not possible to accurately re-create a signal based on insufficient samples or observations. The idea of an energetic imprint is akin to taking a flash photograph of a passing car from a distance of a few millimetres, and trying to infer the make, model and speed of the car form that single datum. It is theoretically highly implausible. Note that science does nto call it impossible, because all kinds of unlikely things are possible, but it’s incredibly unlikely.

Some experiments have suggested something that might be akin to water memory. The first widely reported experiments were by Jacques Benveniste, but these turned out to be the result of (probably unconscious) bias on the part of one of the experimental team, who was emotionally invested in the idea of homeopathy. More recent experiments by Luc Montaigner have shown a short-lived effect analogous to memory but Montaigner said when asked by CBC that you “cannot extrapolate it to products used in homeopathy”.

The law of infinitesimals states that that which causes a symptom can cure it when highly diluted. This contradicts the dose-response relationship as we understand it. It posits an inflexion point in the does-response relationship at which instead of causing the symptoms, it cures them, and becomes more effective in curing them with steadily increasing dilution. There is simply no evidence for this.

New age bollocks

Along the way, homeopathy has picked up a following among New Age types. This has introduced some real gems into the homeopathic repertory such as “venus stella erans” (light of venus) and mobile phone radiation. Rather than giving these people the bum’s rush for dragging down the reputation of the field with patent nonsense, the homeopathy community appears to have embraced them with open arms.

The broken chain

There are a number of links in the evidential chain that appear to be missing or broken for homeopathy.

  • The “law of similars” lacks an evidential basis; there is no provable generalisable mechanism by which something that causes a symptom can be relied on to alleviate the cause of that symptom and it does not appear to be empirically true except for a very small number of well identified cases (see hormesis)..
  • The “law of infinitesimals” breaks down into several problematic assumptions:
  • The idea of water memory is effectively Retrocon, an idea introduced post facto to explain away a conflict between an old idea and new knowledge that says the old idea is wrong.
  • If water memory does exist, is it generalisable? Homeopathic remedies are prepared with everything from dog excrement to the light of venus. Work showing an effect form DNA fragments canot be taken as evidence for the memory imprint of metals or specific frequencies of radiation. Luc Montaigner, heralded by homeopaths as “proving” water memory, says his work cannot be generalised to cover the substances used in homeopathy. That’s bevause he’s a scientist and knows that even if he is right about the effect, it cannot be assumed to be generalised.
  • If water memory does exist and is generalisable, how persistent is it? The effect noted by Montaigner appears to be short-lived, in the nanosecond range, yet homeopaths assert that remedies a hundred years old retain their potency.
  • If water memory does exist, is generalisable, and is persistent, what evidence is there that it also works with alcohol? Many homeopathic remedies are made with alcohol instead.
  • If it exists, is generalisable, and has persistence, you would then have to prove that if dripped on sugar the effective component is transferred to the sugar rather than evaporating with the water. This is a bit of a hard sell. Glucose (left) and water (right) are different in just about every way imaginable; sucrose and lactose (also used) are as complex as glucose.


  • And what evidence is there that this water memory is selective to the point that all the impurities in the water are forgotten and only the desired memory retained?
  • If water memory exists and is generalisable, selective and persistent and can be transferred to sugar, what evidence is there that the effect can survive the process of ingestion? The human digestive system is not a massively complex system but it includes a couple of stages designed to break down complex molecules, absorb certain types of molecule and pass others through without ill-effects. Anything we consume is attacked first by enzymes then by acid, and the broken-down components transfer across the wall of the intestine. At no point is the contents of the stomach or intestine succussed, so no whatever proposed transference mechanism exists to imprint the effect on the original water cannot possibly be in evidence here.
  • If the memory exists, is persistent and selective, can be transferred to intermediaries (without, be it noted, the use of succussion, apparently so necessary during preparation), what evidence is there that any therapeutic effect can thereby be transferred to the human body, still less to the specifically affected parts? Homeopaths prescribe different remedies for the left and right hands, for example. What evidence is there that something taken orally can differentially affect only one of two biologically identical organs?
  • What about mixtures? What about imponderables? Many of the preparations used in homeopathy are not compounds but mixtures, others are made of things like X-rays or the colour blue.

How does it know?

Another problem for the serial dilution process is that water itself contains many impurities. Homeopaths use distilled water, but as every electrochemist knows distilled water only contains fewer impurities, not none at all.

ISO 3696 is a standard for laboratory grade water. Water meeting this standard may not be kept in glass or plastic containers, as they will leach impurities. Glassware for experiments will be washed in hydrofluoric acid, which dissolves the top layer of glass, again to reduce the leaching of impurities. ISO 3696 specifies ten parts per billion of impurities, 10×10-9 – which in homeopathic terminology is 4C. So at the fourth centessimal dilution the impurities in this exceptionally pure and rather expensive laboratory grade water are as significant in the solution as the supposed active principle. There’s no suggestion that succussion can’t affect these impurities the same way it affects the active principle because many if not all of the impurities themselves appear in other homeopathic remedies.

The elephant and the football

Someone once described the idea that molecules leave an imprint on water molecules as being like trying to infer a picture of an elephant by looking at a football after an elephant has trodden on it.

Have a quick look at this:

  • Water molecule

  • Ethanol, also used in preparation of homeopathy

  • Glucose molecule

  • Quinine molecule, fromLife magazine

  • Three forms of DNA molecule

Can you see why science would have issues accepting that DNA and quinine can be transferred in the same way? And why science would not accept without good evidence that water and alcohol behave identically? And why science would have a problem believing that something can transfer from water or alcohol to sugar? Through quantum mechanics we have now a fairly deep understanding of the structure of matter, and pretty much none of what homeopathy suggests is even remotely plausible in the light of this knowledge.

Balance of probabilities

Pharmaceuticals are typically used in adults in doses of the range 50mg to 1000mg; for a 100kg adult male that works out to a few parts per million, a dilution at which some effect is certainly plausible and generally measurable in vitro, sometimes with quite unsophisticated instruments. We do not even have the language to describe the dilutions used in homeopathic products. The term “astronomic” is a common piece of hyperbole and is wholly inadequate to the description of homeopathy – there is not enough matter in the universe to come anywhere close to modelling the higher dilutions. So homeopaths fall back on the language of “subtle energy”, an energy field so subtle that we cannot measure it with instruments than can detect fractional parts of a single atom, yet so powerful that it can dramatically affect health. we seem to be dealing with a vital force of incredible power in relation to its size, whose existence the finest minds have not even begun to contemplate. Scientists have explained or hypothesised everything from the Higgs boson – the so-called “god particle” – to the process of entropy that governs the entire lifespan of the universe; we can draw pictures with single atoms on a silicon substrate, we can split atoms and isolate subatomic particles, we can create antimatter, we can measure the radiation of stars and radio sources hundreds of millions of light-years away, yet we cannot measure the supposed energy field in a homeopathic remedy or the body’s purported energy field with which it interacts, supposedly with sufficient power to cure virtually every disease.

Or maybe it’s just placebo, an effect that is well documented and partially if imperfectly understood.

Before making up your mind which is more likely, you might like to read the Wikipedia article on Occam”s razor.

Appeal to Mystery

Of course some homeopaths say that it’s all a great mystery, and that science does not know anything. It’s true that science does not know everything, but we do know an awful lot and what we do know is consistently inconsistent with homeopathy. The “appeal to mystery” is fallacious; it is arm-waving seeking to dismiss a failure to provide a credible mechanism by which to begin refuting the null hypothesis of placebo effect plus observer bias. If you go to a man who claims he can make rabbits appear from hats by magic, you will naturally believe he is not a magician but a conjurer. He may say that there are great mysteries in the universe but that will not convince you that he has cracked the problem of teleportation when the alternative null hypothesis – sleight of hand – is consistent with everything else we know and teleportation is not.

Unproven claims and patent absurdities

Central to the idea of homeopathy, as with many fields of alternative medicine, is the idea that the body has an energy field. If it does, we can’t detect it with the most sophisticated instruments we have. That does not mean it does not and cannot exist, but the probability of its existence declines as instruments get more sensitive and as we understand more about the universe without finding any supporting evidence for these energy fields. In 1810 perhaps the life force was a mystical thing and it might well be perfectly natural to perceive it as some kind of mystical entity separate from the body itself, albeit that different alt med traditions have different and often mutually contradictory pictures of this force.

The idea that disease is due to some effect of this life-force is much easier to address. The vast majority of diseases can be traced directly to individual pathogens or processes, and the progress and mechanism of disease is generally well understood.

Invalid arguments used by homeopaths

  • Celebrity endorsements are a favourite tool. The implication is that because these people are famous, they are smarter than us or can afford the very best or whatever. It’s pretty obvious that celebrity does not confer immunity from craziness. You wouldn’t take health advice from David Icke or Michael Jackson, so why would you take it from Jerry Hall? Prince Charles supports homeopathy. He also talks to plants. In England, we call this eccentric.
  • Appeal to history or tradition. This is no more valid than the idea of phlogiston.
  • Homeopathy is natural. No it’s not, there is nothing natural about serial dilution and the ingredients, such as arsenic, may well be of natural origin but so are many pharmaceuticals – at root everything comes from the earth somewhere along the line. Being natural does not make something healthy, either: nature is littered with deadly toxins.
  • Homeopathy is licensed. This is not a valid argument because most of the licenses depend on a grandfather clause due to the fact that homoeopathy was extant when regulation was introduced. Most of these regulatoins cite the homeopathic pharmacopoeia, which is a valid source for medical interventions in the same way the Bible is valid as a literal history text.
  • Choice is all very well, society is generally happy to let people choose how to spend their money, but when public money is spent we tend to be a bit more critical and when the products make health claims we are rightly demanding of solid evidence. Medical regulatoin dates back to some very vivid failures in conventional medicine. The FDA can trace its authority to the Elixir sulfanilamide tragedy and many of the UK’s regulations stem from the thalidomide case. As new medicines arrive and replace old, the number of pharmaceuticals which predate the requirement for rigorous testing has dwindled. Your doctor no longer has the choice of prescribing outmoded or dangerous drugs, and that is as it should be. Homeopathy has been given a free pass on this largely because it is harmless – there is no credible evidence that homeopathic remedies contain any pharmacologically active compound.
  • Science doesn’t know everything is a statement of truth and probably always will be, but science does know that there is no credible mechanism by which an effect can be caused by a substance diluted beyond the point that it no longer exists in the solution.
  • Drugs are toxic is also in some cases a statement of truth, especially in the field of cancer medicine where the tools at our disposal are in some cases deliberately toxic because there is a need to kill the body’s own cells. It is not, however, a valid criticism of the idea of modern medicine in its entirety, nor even in the specific case of toxic chemotherapy. An evidence-based treatment has to be shown to be effective, where it is not, or where the harm can be shown to outweigh the good, it fails. We no longer give live poliomyelitis vaccine because polio has been largely eliminated so the less effective dead vaccine, which has a lower chance of undesirable side effects, is adequate to maintain herd immunity.
  • Drugs kill and injure x hundred thousand people a year is a variant on the above and often relies on some creative definitions of harm (such as accepting at face value the anti-vaccination movement‘s claims of vaccine damage, none of which stand up to independent scrutiny. But even if it were true, UNICEF say that nine million lives annually are saved by vaccination alone, a figure that dwarfs even the most pessimistic claims for medical harm.
  • It works for me is a fallacious argument: evidence is not the plural of anecdote, and the number of cases of health improvement that can be unambiguously and objectively attributed to a homeopathic remedy, rather than placebo effects, is precisely zero.
  • It works for millions in India is a fallacious argument, argumentum ad populum, and in any case who says it works? India’s average life expectancy is a decade shorter than that of most Western countries.
  • People who hate homeopathy are big pharma shills is invalid on three separate grounds. First, in general people don’t hate homeopathy, they merely deride it as modern-day snake oil (though they may well be angry at those who promote it in lieu of effective interventions such as antimalarials). Second, most sceptics are not connected in any way with so-called “big pharma”. Third, it assumes there’s no such thing as “big alt-med” – Boiron is a multi-million-dollar corporation which spends nearly twenty times as much on marketing as R&D, whereas most pharmaceutical companies spend 2x or less.
  • Various conspiracy theories are also advanced to explain away opposition, Government support for vaccination (in the US the Government had to step in because vexatious lawsuits were undermining the not terribly profitable manufacture of vaccines to the point that only one company was still producing some vaccines). Conspiracy theories are virtually impossible to argue against. As far as conspiracists are concerned, the more you prove that there is no evidence to support them, the more conclusive the conclusion that all evidence is being suppressed. Conspiracies around vaccination and the like belong in the same box as alien abductions, Elvis and the like, with very rare exceptions (such as the case of Andrew Wakefield, who it is safe to conclude probably was working to advance vested interests, albeit quite likely from the purest of motives; a very sad case).
  • The proving process is a form of clinical trial! Well, no, actually it’s not. A proper clinical trial is randomised and double blinded – neither the patient nor the doctor knows whether the patient is receiving the test treatment, a placebo, or perhaps a comparison treatment of known effect. There have been some interesting studies around this, including in the trials ofantiretroviral drugs in the US whose development was hampered by collusion between patients, who pooled and split the medication reasoning that it would be better for everyone to have a half dose than for half of patients to receive on treatment. In a homeopathic proving a substance is given to a very small number of people, who are almost never blinded, by a researcher who is almost never blinded, and the patients are asked to record the symptoms. In many cases there are so many symptoms of such vagueness that the conclusion of a sceptical observer would be that the preparation causes hypochondria.
  • Many studies support effect is true, but not relevant; these studies are often of very poor quality and there are very good reasons why most published research findings are wrong. The best way of assessing a disparate evidence base is with well-designed systematic reviews. At least five of these have been done and the conclusions are (a) that homeopathy does not appear to have any effetc beyond placebo and (b) the more rigorous the design and conduct of a study, the more likely it is to report no benefit above placebo.
  • Nobel scientist supports homeopathy! is actually false – the scientist in question says his work cannot be extended to homeopathy and in any case “water memory” is only one of a long list of broken links in the chain of evidence.
  • Succussion generates nano-bubbles – or, to translate into valid scientific terms, shaking causes bubbles. Who knew? Putting nano in front of something does not make it special or mahic.
  • It’s a quantum effect is simple technobabble, it has as much meaning as dilithium crystals.

Ported from my Wiki.


  1. John Monroe

    In your Venn diagram, the homeopathy circles should be completely inside the woo circle. What part of homeopathy is not woo? A part of medicine as practiced in the 19th century was also woo. A small part of modern medicine, if you define it to include all treatments by medical doctors or registered nurses, in accredited hospitals, or paid for by health insurance is unfortunately also woo.

  2. John Monroe

    In your Venn diagram, the homeopathy circles should be completely inside the woo circle. What part of homeopathy is not woo? A part of medicine as practiced in the 19th century was also woo. A small part of modern medicine, if you define it to include all treatments by medical doctors or registered nurses, in accredited hospitals, or paid for by health insurance is unfortunately also woo.


    • The therapeutic value of sympathetic listening, and not using medicine to fix life simply sucking.


  3. Thank you for the great article, but please, next time use spell check and get a good proof reader.


    • Or maybe go back I time and unburn my left hand so it doesn’t have bone deep scars across the knuckles.


Leave a Reply