Due to a disk crash and backup failure, this site has been restored from an old backup with a number of more recent articles missing. The missing site content is being restored as time permits. We apologise for any inconvenience.

In Defence of Feyerabend

A Constrained Rant by The Famous Brett Watson, 16-Jun-2003.


Paul Feyerabend stands somewhat apart in the annals of philosophy of science. Whereas most philosophers of science have attempted to say what the scientific method is, Feyerabend's most prominent contributions in the field have denied that there even is such a thing as "the scientific method". Likewise, where most other philosophers have opted for formalism (and sometimes formalism piled upon formalism), Feyerabend stands out from the crowd as a provocative employer of rhetoric, mowing down the tall poppies (or Poppers) of philosophy with refreshing rascality. His watershed publication in this area was the book, "Against Method", and it earned him a great deal of ire from the philosophical establishment. Much of Feyerabend's style, and the severity of the backlash he experienced from publishing Against Method, can be gleaned from a single sentence in his autobiography, where he says of it, "I often wished I had never written that fucking book."[1]

To appreciate Feyerabend's point of view, it is important to be aware of the context into which his ideas were introduced, with particular reference to the prior work of his contemporary, Thomas Kuhn[2]; namely, The Structure of Scientific Revolutions. Kuhn's work was most notable in that it looked at science from a largely historical perspective, rather than trying to present a formal description of science as a process. His conclusion was that science goes through periods of normality, punctuated by periods of radical revolution ("paradigm shifts"), in which the whole perception of that which is being studied changes. Such revolutions more or less precluded the possibility of a single scientific method, since methods and meanings could change incommensurably between paradigms. Some critics saw in this model an implication that science was not a rational process. Broadly speaking, Kuhn distanced himself from that implication, and Feyerabend embraced it.

In this essay, I will evaluate (and largely defend) some aspects of Feyerabend's paper, How to Defend Society Against Science[3], which dates from about the same time as the publication of Against Method. In it, Feyerabend argues that there is no special, rational, scientific method, and concludes that science does not deserve the exalted status that it enjoys in today's society; indeed, that science can be viewed as a menace against which society ought to be protected. This idea seems prima facie utterly preposterous to many people; as preposterous as the idea that we would be better off without modern medicine, or automation, electrical appliances, and so on. On closer investigation, however, I believe that Feyerabend's paper can be found to contain vital truths, so bear with me.

Naïve Dismissals

Before embarking on a constructive criticism, I feel it best to deal with some of the relatively naïve arguments which might be used to dismiss Feyerabend out of hand. Feyerabend is uncommonly brusque in this paper; he addresses a radical message to a conservative audience with intent to ruffle some feathers. The less sophisticated of his detractors are often curt in response, but fail to match (or recognise) the rigour of his argument. I will here briefly set up some straw-man opponents for Feyerabend, then knock the stuffing out of them so that we aren't distracted by them during the constructive part of the critique.

Our first straw-man believes we should ignore Feyerabend because science works. If the progress of science weren't self-evident enough, then the accompanying march of technology must surely be proof. When did any religion offer us such technological progress? Surely, therefore, one cannot construct a meaningful case against science as knowledge; the only kind of argument in which one can reasonably engage is that of why science does work.

This straw-man is largely playing definitional games; equating "science" with "progress". Under this definition, something is "science" to the extent that it effects "progress", for values of that term principally relating to technology. But note that this formulates science in terms of its results rather than its methods, and Feyerabend's main thrust is against method, not against results. Thus, this argument would appear to be a non sequitur; it fails to address Feyerabend's actual point. Even if we were to grant the very generous assumption that the results in question can be best obtained by a particular "scientific method", the selection of results is value-laden. On what rational grounds could we say that the person who prefers "spiritual progress" over "technological progress" is wrong?

To this, our straw-man may retort that technological progress is objectively measurable, whereas anything that might be called "spiritual progress" (whatever it is) is subjective at best, and purely imaginary at worst. This is, of course, absolutely true, so long as we accept the metrics used in making the judgment. But the preference for one system of metrics over another is also a value judgment, and therein lies a strong indicator that science could be something against which society ought to be defended. No doubt our straw-man considers it bad for religions to impose their value-judgments on society as a whole; why should it be different for science?

Our second straw-man thinks that Feyerabend is misguided because science is progressive, and better than the alternatives. Science is not perfect — far from it, in fact — but it is vastly superior to its alternatives (again, as evinced by scientific progress). Even if there are alternative theories, we should not waste time with them in education precisely because they are already known to be inferior. Education is a finite resource, and there are an infinite number of bad theories. The only thing that ought to replace the best current theory is a better one, not a worse one.

The position held by this straw-man is similar to that of the first, but more patronising. It contains the same set of value judgments, by which some particular formulation of "science" is judged as superior to its alternatives, plus the additional judgment that only the best theory ought to be presented in education even if we acknowledge that "best" is not "perfect"; that positions which have been judged as inferior should not be allowed to waste precious time. Despite the moderate phrases in which our second straw man expresses his case, this objection is less reasonable than the first: not only would he impose on us his values with regards to progress, but also with regards to education. Feyerabend takes a directly opposing view with regards to the exclusion of contrary ideas because he thinks it results in better education. In a few words, one could object to our second straw-man on the grounds that he is advocating indoctrination, which is surely something he would object to if a religion (or other disagreeable party) imposed it.

Our final straw-man says that Feyerabend should be dismissed because his intellectual anarchy would be worse that what we have now. It would herald a return to the pre-scientific dark ages. Regardless of whether there is any technical merit to Feyerabend's thesis against method, we should reject it on pragmatic grounds, always preferring whatever method is in vogue at the time to the absence of method.

I believe that this straw-man has simply misread Feyerabend. I believe that Feyerabend does offer us a method, or meta-method, and that this method ought to be attractive to many hard-headed devotees of science (in its various forms). This matter forms the subject of the remainder of the essay.

The Positive View

Feyerabend's positive contribution in How to Defend Society Against Science is spelled out in the section entitled "Against Method".

There is only one view which overcomes all these difficulties. It was invented twice in the nineteenth century, by Mill, in his immortal essay On Liberty, and by some Darwinists who extended Darwinism to the battle of ideas. This view takes the bull by the horns: theories cannot be justified and their excellence cannot be shown without reference to other theories.

— Paul Feyerabend, How to Defend Society Against Science


At the time I first read this, I knew of Mill's On Liberty, but had not actually read it, so it was not entirely clear how much of Feyerabend's following commentary related to it, and in what way. Clearly, though, if we are to take Feyerabend at face value here, we should understand that he believes Mill to have already solved the problem of science, and we cannot hope to properly understand his argument without reference to Mill. Thus, I turned to On Liberty, and found the relevant part to be chapter two, Of the Liberty of Thought and Discussion. In this, Mill puts forward the thesis that we should not only allow, but encourage a diversity of opinion, and he gives three major reasons for this stance.

Firstly, opinions ought not to be stifled on the grounds that we consider them false, for we ought not to presume ourselves to be infallible judges. Even in science, where the real world is supposed to act as an arbiter of opinion (or theory), this is obviously true. Some would consider the history of science to be a progression towards "greater truths", but the idea is problematic when spelled out in any detail. We certainly get the impression of general progress, particularly with regards to technology, and one might reasonably suppose this to reflect progress towards a more accurate understanding of reality itself; but on the other hand, the strongest trend in science has been the inevitable discovery of flaws in incumbent theories which lead to their substantial revision or replacement. The incumbent theory may currently have the weight of evidence, but this does not make it true and its alternatives false. Mill's argument seems, thus far, entirely applicable to science.

Secondly, even if we were somehow sure of the truth of our opinions (or theories, in the case of science), and the falsehood (or inferiority) of the alternatives, we ought not to stifle the alternatives, for the following reason.

However unwillingly a person who has a strong opinion may admit the possibility that his opinion may be false, he ought to be moved by the consideration that however true it may be, if it is not fully, frequently, and fearlessly discussed, it will be held as a dead dogma, not a living truth.

— John Stuart Mill, On Liberty

This second point is what Feyerabend emphasises when he addresses matters of education, and why he talks of science in terms of dogma. Those who think that science ought to occupy an exalted seat in the pantheon of knowledge typically hold this position because science can in some way be distinguished from mere dogma. The nature of this distinction can take many forms, such as "empirical evidence", "falsifiability", "predictive ability", and so on; perhaps the common thread is some sort of grounds in tangible reality. Thus, science is not dogma, and Feyerabend is just plain wrong. Be that as it may, science can be taught as dogma, and this is what Feyerabend cites as problematic in modern education.

When teaching a myth we want to increase the chance that it will be understood (i.e. no puzzlement about any feature of the myth), believed, and accepted. This does not do any harm when the myth is counterbalanced by other myths: even the most dedicated (i.e. totalitarian) instructor in a certain version of Christianity cannot prevent his pupils from getting in touch with Buddhists, Jews and other disreputable people. It is very different in the case of science, or of rationalism where the field is almost completely dominated by the believers.

— Paul Feyerabend, How to Defend Society Against Science (emphasis in original)

The problem here, as Feyerabend sees it (based on Mill), is that the absence of contrary opinions necessarily turns science (or anything else) into dogma. An idea is not dogma only when it faces vigorous challenges from alternative ideas. A defender of science may interject here, claiming that it is, rather, the grounds in tangible reality that makes science non-dogmatic (and religions otherwise). I disagree: it is the grounds in tangible reality that makes science science (and other things non-science), but science is in peril of dogma in two senses. The first is that particular scientific theories can be held dogmatically, and it might be said that Kuhn's periods of "normal science" (between revolutions) are those periods in which a particular scientific dogma prevails. Secondly, the idea that "scientific knowledge deserves special status" is itself a dogma unless it is held in the face of serious competition from opposing views.

Mill's argument is quite radical; it poses "dissent" as a vital ingredient in human understanding. If we are only acquainted with one side of a story, then we know less of it than we would if we were also familiar with the other. Mill goes so far as to say, "So essential is this discipline to a real understanding of moral and human subjects, that if opponents of all important truths do not exist, it is indispensable to imagine them and supply them with the strongest arguments which the most skilful devil's advocate can conjure up." That is, if we can't find dissent, we ought to manufacture it! Is science exempt from this? Feyerabend thinks not, and I believe the idea should give all philosophers of science pause, since the very idea of manufacturing dissent runs in such stark contrast to all attempts to formalise the scientific process.

The third point that Mill raises is less controversial than the one just examined, but quite worthy of note; namely, that rarely (if ever) in human arguments can either party be said to be completely right or completely wrong. This would also appear to be uncontroversially true of scientific arguments. The wave/particle duality of light provides a simple example: in times past it was argued that light had a particulate nature; this was overthrown by wave theories; it's now recognised that neither model is complete and correct. The moral of the story is that we shouldn't consider older theories to be completely wrong any more than we should consider current theories to be completely right. The old theories must have had some evidential support, and science can be expected to exhibit some "cycles of fashion" as a consequence of this fact.


Although Feyerabend does not cite any particular source when he mentions the application of Darwinism to "the battle of ideas", I think the relevance of the issue is fairly obvious. If the competitive forces of evolution can give us the very brains by which we are able to conduct this discussion, then competitive forces between opposing views must surely work to the advantage of those views in terms of weeding out weak arguments, fallacies, anomalies, and so on. Mill's suggestion that we should manufacture dissent is the intellectual equivalent of a fighter seeking a competent sparring partner. We typically acknowledge "competition" as an agent of improvement in biology, economics, and sports; why not, then, in the field of science?

A defender of science might object, at this point, that we do have such competition in the field of science. There are, for example, competing theories with regards to the mechanism of evolution; and scientific revolutions could not occur at all unless there were at least two competing theories at the time. These points are true, but the diversity exists despite the prevailing culture, not because of it. Diversity is not encouraged; the existence of competing theories is viewed as a problem, since at least one of them must be false. Feyerabend addresses this when he compares Mill's approach to that of Karl Popper.

Finally, Popper's standards eliminate competitors once and for all: theories that are either not falsifiable or falsifiable and falsified have no place in science. Popper's criteria are clear, unambiguous, precisely formulated; Mill's criteria are not. This would be an advantage if science itself were clear, unambiguous, and precisely formulated. Fortunately, it is not.

— Paul Feyerabend, How to Defend Society Against Science

I believe a sporting metaphor illustrates the difference between current practice and Mill's prescription quite well. The existing field of science resembles a competition in which there exists a single reigning champion; challengers must face the champion, and the winner claims the title. Philosophers of science observe the competition, and attempt to discern the technique that the champion employs to his advantage, or the novelties of technique which allow an upstart to claim the title. Mill, on the other hand, would have us engage in a tournament of many teams, competing in a round robin (rather than knockout) format. Teams which are generally considered to be weaker are not to be eliminated (although they may spontaneously disband), but are retained because the competition benefits all players. Even a weak team may score some points against a strong one.

Our second straw-man may return to reiterate that we can't possibly support an infinite number of alternative methods. I agree, but I will insist (with Mill and Feyerabend) that the ideal number of competing methods is greater than one. I would consider it good progress if even one distinctly unpopular viewpoint were allowed representation where it is currently shunned. I think the ideal situation would be to have as many competing methods as there are teams prepared to back them; we are not so over-blessed with ideas that they outnumber participants — far from it. Our straw-man is warning against an extreme state of affairs, when it is, in fact, the other extreme which exists.

That the ideal number of competitors is greater than one seems almost self-evident. In what way can there be competition with only one competitor? That an absence of competition is harmful is perhaps most obvious in the economic realm, where it takes the form of monopolies. It is the natural tendency of monopolies to be unprogressive, since their aim is to maintain the status quo, a goal often more easily effected by causing direct harm to their competitors than by out-performing them; i.e., they behave anti-competitively. The same "anti-competitive" behaviour is likely to find its way into any school of thought which gains too much of an upper hand in society; thus Feyerabend's suggestion that science is not inherently liberating, even if it was a liberator in the seventeenth and eighteenth centuries. Modern scientific/rationalist thought has an effective monopoly, and this isn't good for people in general, or for the progress of science itself.

Strangely, the application of Darwinism to the sphere of ideas doesn't seem to be held in nearly as high regard (by Darwinists) as one might expect. In his 1976 book, The Selfish Gene, Richard Dawkins coined the term meme to describe the conceptual equivalent of the gene in the realm of ideas[4], and I hesitate to describe Mill's prescription as "natural selection of memes" only because the term seems to have acquired a negative connotation. To go into detail here would be tangential, so suffice it to say that memes are frequently depicted as nasty parasites which "infect" our thinking, making us behave in irrational ways; the phrase, "that's just a meme," carries the connotation that the idea is contagious but somehow meritless. Implicit in this connotation is the assumption that some ideas ought to be exterminated, and this runs contrary to Mill.


Feyerabend, with reference to Mill and Darwinism, provides us with a recipe for scientific progress, or at least a recipe for an environment which encourages it. This is distinctly different from what other philosophers of science have attempted to provide; namely, a formal account of "the scientific method". In closing, I would like to briefly discuss this distinction, and argue as to why Feyerabend's approach is the better one.

It is a common mistake to think of Feyerabend as "anti-science". He is only anti-science to the extent that he is pro-freedom, and sees science as a tyrant. He does not claim that science is dogma, but rather that science has become dogmatic, as does any ideology which gains an effective monopoly. Feyerabend supports liberty of thought, and this puts him at odds with those who insist that scientific reasoning is the superior mode of thought. There is a deep irony in the fact that many who would call themselves "freethinkers" are now part of a movement with sufficient clout to be a source of intellectual oppression.

Science is not the only worthwhile human goal, and within science as a goal there is no one proper method. Perhaps, if the task of science is ever completed, we will be able to look back over its history and discuss whether any particular method would have been sufficient to the entire task. In the meantime, we are better off with many methods than with dogmatic adherence to any single method. So let us have many methods, and many spirited debates as to why one method is better than another. Let the practitioners of each method boast with their results, the progress that they make, the technologies they develop, the discoveries they bring to light, their explanatory or predictive power; and let them adopt all the best techniques of their opponents as they recognise them.

It may well be that science deserves an exalted seat in the pantheon of knowledge, but, on the other hand, maybe it is no more or less important than other kinds of knowledge. Whatever the case, if Feyerabend and Mill are right, all branches of knowledge should adopt an attitude of humility, and encourage diversity of opinion rather than engaging in a process of elimination. They should do this for the sake of their own progress, as well as for intellectual liberty in general.


  1. Preston, John, "Paul Feyerabend", The Stanford Encyclopedia of Philosophy (Summer 2002 Edition), Edward N. Zalta (ed.), <http://plato.stanford.edu/archives/sum2002/entries/feyerabend/>
  2. Contemporary indeed: Feyerabend and Kuhn were both born in the 1920s and died in the 1990s.
  3. Originally published in Radical Philosophy #11, Summer 1975, and itself a "revised version of a talk given to the Philosophy Society at Sussex University in November 1974".
  4. Memes are the subject of chapter 11 of The Selfish Gene.


Dawkins, Richard, The Selfish Gene, Oxford University Press, 1976

Feyerabend, Paul, How to Defend Society Against Science, Radical Philosophy #11, Summer 1975

Mill, John Stuart, On Liberty, 1859

Zalta, Edward N. (ed.), The Stanford Encyclopedia of Philosophy (Summer 2002 Edition)

Nutters.org Author: The Famous Brett Watson
Date: 2003-06-16
Public Domain: the author waives copyright on this document. Other sources (if any) are quoted with permission or on the principle of "fair dealing" and retain their original copyrights.