< back

by Ian Kim

Fayerabend: The Last Savior of Science


Critics of Feyerabend all follow patterns of behavior undeniably similar to each other. No matter which way they turn, a simple extension of their logic regresses to the beginning of Feyerabend’s argument. In no other criticism is this more apparent than in Chalmers’ “middle lane” counterargument. When drawn to its conclusions, we see that Chalmers puts himself in a double bind. If he wishes to end anarchism, Chalmers must make dogmatic claims which stand at the beginning of Feyerabend’s normative arguments. Otherwise, Chalmers must succumb to relativism. This pattern also reveals the importance of remembering, embodying, and continuing Feyerabend’s cynical and lighthearted criticism of science. In the face of unreasonable effectiveness, it is entirely too easy to fall into dogma. We must continually embrace the paradoxical, the illogical, and the humorous to guard the integrity of science against the very individuals who claim to support it.

Epistemological Anarchism

From the opinions of the Vienna circle and the logical positivists, we find the scientific method beaten and defeated by the attacks of Popper, Duhem, and Kuhn. First, they took away induction as an instrument of gaining knowledge, then falsificationism, then even our theories were deemed incommensurable. By the time Feyerabend launched his offensive, the groundwork was already laid for the defeat of science.

Feyerabend’s offensive in Against Method is elegant in form and simple in character. His epistemic argument simply contends that the world is a complex place. In the face of unfathomable complexity and unknowness, it simply does not make sense to limit ourselves to a particular method. We should instead “keep our options open, and we must not restrict ourselves in advance” (Feyerabend, 20). Additionally, how ignorant would it be to believe that such a universal method can discover any underlying mechanisms of nature? In the concise history of science, all proposed “universal” methods were shown to be useless, inconsistent, and counter-productive. The only ‘rule’ that survived was “anything goes” (Feyerabend, 296).

Given science’s presumptive and dogmatic state, the perceived differences between science and religion get smaller and smaller. How can science claim higher status when its participants no longer view themselves skeptically but with the dogma of ‘primitive’ religions (Feyerabend, 301)? If Feyerabend is correct about the core characteristics of religion and science, it truly deserves no higher status. Since the American government values the separation of church and state, there should also be a separation of science and state. Why teach science as truth over myths when we do not teach Christianity as truth over Islam? Feyerabend, having made his epistemic argument, also launches a political offensive. Man’s flourishing can only be achieved by exploring his own interests, projects, and activities. However, scientific education in all its dogma compresses every part of our individuality “like a Chinese lady’s foot” (Feyerabend, 20).

Methodological Change in Method

Chalmers is very much sympathetic towards Feyerabend’s campaign, but only insofar as it applies to universal ahistoric accounts of method. For example, Popper proposed his falsification schema as a timeless, boundless, universal principle for making progress in science. According to Popper (and those who have tried to formulate the method after him), the same standard of falsification should be applied to vindicate Aristotelian physics as well as modern quantum mechanics. Chalmers, like Feyerabend, believes that philosophy has neither the warrant nor resources to make definitive claims about what the appropriate standards of future sciences should be (Chalmers, 150). As it appears, no such formulation can stand the test of time.

Chalmers instead proposes an alternative: if science is an open-ended quest for knowledge, why can’t we adapt and improve our methods and standards as we learn more (Chalmers, 150). This so-called “middle way” rejects the existence of a universal method while also rejecting “anything goes.” The immediate worry, of course, is whether one can successfully define a methodological change in the method without falling into the same traps that Popper, Kuhn, and Lakatos have. To stay in the middle, Chalmers must give some sort of overarching value or standard that methods converge towards (Chalmers, 150). However, if these overarching values and standards exist, we have simply bubbled the same problem one level up. In the context of unfathomable complexity, how can we rely on these overarching values and standards to guide us the right way? These values are still just as dogmatic and ignorant as the methods put forth by previous philosophers. Philosophers of science such as Worrall are familiar with these criticisms. Nevertheless, Chalmers quickly demonstrates how changes in method can occur without appealing to overarching super values.


In order to refute the viability of Chalmers’ model, I present the following propositions. First, through Chalmers’ portrayal of Galileo, we see that no matter what Galileo does, changing the opinions of others relies on a certain leap of faith that cannot be made logically. Second, even if Chalmers can establish that no leaps of faith are necessary in Galileo’s case, he presents no generalizable argument, simply an example to the contrary (only to face a litany of Feyerabend’s other examples). Third, given the necessity of leaps of faith, we see that Chalmers’ methodological change in methods is simply a description of anarchism in hindsight. Therefore, in order to truly argue for the “middle lane,” Chalmers must become dogmatic.

The Portrayal of Galileo from Both Sides

Chalmers and Feyerabend propose two competing depictions of Galileo. To Feyerabend, this poster child of scientific triumph uses rhetoric and epistemological tricks to exploit opportunities shamelessly. To Chalmers, Galileo finds common ground with his opponents and uses simple irrefutable demonstrations to convince and persuade. The central point of contention lies around the issue of instruments in science. For Galileo to be taken seriously, he first had to convince others that telescopes were a valid instrument of science.

Of course, for terrestrial use, it was straightforward to convince people. Simply point the telescope at a distant object, then travel to the distant object to check the observation yourself. However, this was impossible for heavenly objects (Chalmers, 142). Galileo simply could not travel to mars or venus. We also know how to adjust for specular distortions and other “tricks” instruments play on our senses because we have access to the very thing we are observing, but the same cannot be said for heavenly bodies. So, according to Feyerabend, Galileo resorted to trickery and rhetoric. Instead of evidence, Galileo used “clever persuasion techniques, because he wrote in Italian rather than Latin” and appealed to those eager to accept new ideas (Chalmers, 143). For Feyerabend, circumstances, language, emotions, and even mystical faith take precedence over an argument’s rational evaluation, even when they are available. Chalmers paints a very different picture by suggesting that despite the overwhelming preference for raw senses, Galileo found common ground and built up his argument from small irresistible claims to convince his opponents ultimately.

The Aristotelian preference for raw senses dominated the scientific view: If nature creates everything for a purpose, and if the purpose of man is to understand nature, then nature surely would not make our senses inherently unreliable (Chalmers, 151). In order to overturn this standard, Galileo appealed to irradiation (Chalmers, 152). When bright stars are viewed against the night sky, objects appear much larger and “festooned with adventitious and alien rays” similar to how candlelight appears larger when farther away in the dark. Since Galileo formulated the appearance of an object as a function of brightness, size, and distance, he did not have to appeal to telescopic data at all. For example, the brightness can be reduced by viewing through “a cloud, a black veil, colored glass… etc.,” to mitigate irradiation (Chalmers, 153). These formulations can also be directly tested on earth. Simply move a light source from the subject into the dark, and you can clearly observe the apparent size change.

Galileo could also extend this principle to heavenly bodies. The apparent size of Venus, for instance, is much larger during the night against a dark background than during the daytime (Chalmers, 154). So then Galileo had a viable argument for the existence of irradiation for both terrestrial and heavenly objects, which could all plainly be demonstrated. This, combined with the observation that telescopes decrease irradiation, provides an excellent foundation for the veracity of telescopic data. Galileo’s final blow is that telescopic data on the apparent sizes of Venus and Mars align precisely with the predictions of other competing astronomical theories at the time.

Chalmers’ Tricks

Imagine yourself as an opponent of Galileo. Our raw senses (given that you are not a habitual drunkard) take supremacy over instruments because nature and God willed it so. This “pessimist” may accept the phenomenon of irradiation in terrestrial and celestial contexts. This “pessimist” may also accept that the telescope limits irradiation in terrestrial contexts. However, these propositions cannot directly conclude that the telescope provides reliable observational data for heavenly objects.

Imagine that you are in a room with a blocked-out window and a virtual reality goggle. This headset is a black box; you do not understand how it fundamentally operates. However, when you put them on, you see the same room. You are given unlimited time to check the correlation between headset reality and reality. It appears as though the contents of the headset are a virtual replica of the real world and all its dynamics except for the window. You see a tree (which still abides by observed physics phenomena of the room) when you peer out the blocked-out window with the goggles. Of course, one cannot blame this individual for accepting the headset data as it pertains to observing objects in the room. However, there are no deductively valid reasons to “force” this individual to believe that the tree should be real because there is a level of unmendable distance between reality and headset content. Is it not entirely possible that the underlying algorithm used to represent the real world virtually has some flaw that may lead to differences for what lies beyond the blocked-out window? Or even that the tree seen from the googles is entirely made up? Until we can verify each line of code and circuit that compose the headset, we cannot logically trust observations gained from it. As long as you cannot ground-up understand the mechanisms of the headset, reality and headset-reality will remain separate from each other.

Similarly, as long as you cannot ground-up understand atmospheric dynamics, real appearance of heavenly bodies, and optics, the telescope-reality will remain separate from reality. As long as there is a distance between trusted and untrusted input, what is required is a leap of faith. Now, I concede that Galileo may have made that leap a tiny one to take for his peers. Nevertheless, no matter the evidence, he is condemned to fail in completely bridging the gap. There were no deductively valid reasons to force individuals to accept telescopic data; only a leap of faith. Chalmers tries to address this by appealing to previous theories. If the data on the apparent sizes of Venus and Mars align precisely with the predictions of competing astronomical theories, surely telescopic data should be trusted.

Chalmers lacks the big picture. If Galileo proved the veracity of telescopic data by appealing to alignment with competing astronomical theories, which he sought to overturn in the first place, that is not a piece of evidence that Galileo can use either way. Galileo cannot use the veracity of competing theories (Ptolemaic astronomy) to justify both the instrument itself (telescope) and the theory supported by the instrument (Galilean astronomy), which contradicts the competing theories used to justify the instrument in the first place. Even in Chalmers’ depiction of Galileo, he uses blatantly fallacious but convincing ‘tricks’ to nudge people off the edge. Hence, we find ourselves back at square one. At the core of the significant shift to instrumental data was not an airtight deduction but simply a leap of faith. Galileo brought his opponents to the edge through logical and fallacious arguments, but the choice to leap was entirely up to the individuals’ subjective beliefs, values, and history.

Chalmers also only refutes the protective edge of Feyerabend’s arguments, not the hardcore. Feyerabend could simply abandon his Galileo example, and Chalmers would have to face a litany of other examples. Chalmers’ criticisms are not generalizable either. For example, there is no common ground between the pendulum and falling rock interpretations of pendulums because they are both correct from the beginning. There is no “venus” to see in broad daylight. Perhaps one could make a very compelling argument that one should view the pendulum as a pendulum, but no such argument can ever force the pendulum view. That leap of faith can only be made personally and subjectively. So Chalmers pretends to have stripped arbitrary factors out of scientific persuasion, only to be left with a core of personal and subjective factors which actually push people off the edge.

Chalmers: an Anarchist in Disguise

In the end, the model argued by Chalmers is not a methodological change in method but simply a description of methodological anarchism. We must differentiate between methodology at the personal, discipline, and science levels to understand this criticism. What Kuhn, Lakatos, and Popper tried to establish is a method at the discipline or science level: a set of values, standards, and methods that can be applied to all science at all points in time. On the other hand, we have personal methods: a set of foundational beliefs such as one's goals, benchmarks, and rough paths to achieve results held by individual researchers. Notice that the existence of personal methods is entirely consistent with Feyerabend’s epistemological anarchism since one’s individual goals, benchmarks, and postulated paths are all subjectively determined and can be changed at will by the personal persuasions of the individual. The individual is free to adapt, modify, and utilize his method as he sees fit. He does not possess the warrant nor the authority to lecture others on the benefits of his method that could logically ‘force’ others into belief. Others can only make a certain leap of faith into it. What we are left with is a stricter epistemological formulation of the non-aggression-principle: any logically forceful subjugation and conversion of one’s subjective beliefs is simply not possible.

What Chalmers sees when he talks about the progressive modification of “any part of the web of aims, methods, standards, theories, and observational facts” is simply a representation of methodological anarchism (Chalmers, 157). An anarchist society is not one where all beliefs lie on a uniform distribution. Like any other metric across any society, the distribution of personal methods will result in a non-uniform distribution. The web Chalmers sees is simply the dominant aims, methods, and standards of personal beliefs held by each researcher. Of course, when you look at it from a macro perspective, it seems like there are progressive methodological modifications to method. However, when you examine deeper, you see that there are only a collection of personal beliefs, which can only be changed through subjective leaps of faith (no matter how small the leap is).

Boyle’s mechanical philosophy and its abandonment is a prime example of Chalmers’ confusion. Boyle’s theory of matter and corpuscles required all explanations of physical processes to be traced back to the primary motion of corpuscles (Chalmers, 156). Having failed to meet these unreasonable standards, Boyle nevertheless charged forward and performed experiments in pneumatics and chemistry, which yielded results that stuck around (Chalmers, 157). Boyle’s mechanical philosophy was his own personal method, born out of subjective, arbitrary factors and circumstances, which was abandoned for experimental results due to arbitrary factors and circumstances. No other scientist was ‘forced’ due to the undeniable deductive nature of Boye’s philosophy or its revision at any point in time. How can the example of Boyle’s mechanical philosophy and its abandonment serve as evidence for an incrementally changing ‘method’ if only Boyle bought into it in the first place? Mechanical philosophy was simply Boyle’s method, not the method which later changed.


Chalmers’ arguments are therefore incomplete. If he only wishes to make a descriptive claim about the macro-structure of scientific ‘progress’ as a shift in the web of “dominant” aims, methods, and standards where the underlying force of change is leaps of faith and subjective belief, then he is in complete agreement with methodological anarchism. If scientists are allowed to make subjective judgments about their personal methods without ever being logically forced into changing their minds, then that is the definition of “anything goes.” If he wishes to argue against methodological anarchism and truly take the “middle road” approach, he must make an additional commitment that the ever-changing web should become every scientists’ aim, method, and standard. However, Chalmers refuses to make this claim because then we end up back in square one. Feyerabend’s normative arguments begin with the accusation that dogmatic scientists foolishly enforce their dogma on everyone else. But would the method not change under Chalmers’ model? Well, of course, but the tenets of Catholocism are still dogma despite Pope Francis’ countless revisions and reinterpretations. Dogma is not purely dogma because it is immutable, but because people are told to believe in it. Chalmers is therefore stuck between a rock and a hard place. If he wishes to refute Feyerabend, he must embody the very thing Feyerabend accuses him of being. If he does not embody the dogmatic scientist, he has no choice but to agree with Feyerabend.

Herein lies the beauty of Feyerabend’s work. Feyerabend revels in paradoxes and contradictions and yields them as his weapons of choice in the face of serious men stuck in their logic driving themselves into corners. Of course, Feyerabend argues for methodological anarchism using logic, but the reader understands (or should understand) that Feyerabend is ultimately just entertaining himself. It is easy to lose yourself in science’s unreasonable effectiveness and power. Once you see the marvels of modern technology, one begins to think, “this is so great; why doesn’t everybody do this.” In this intoxicating moment, we often forget that progress is made through expressions of individuality, subjectivity, and sometimes faiths in witchcraft and nonsense. Humor and unseriousness is the solution to this intoxication. By being boisterous, provocative, and contradictory, Feyerabend reminds us that none of this should be taken too seriously. Seriousness begets dogma, dogma begets indoctrination, and indoctrination begets compression of individuality which cripples scientific progress. Paradoxically, by savagely tearing down and doubting science, we are protecting its hard core from those who claim to preserve it.