Popper Without “Refutation”

By Bruce Nielson

Full Presentation Available at The Theory of Anything Podcast, Episode 41

Thomas Kuhn’s Challenge to Popper

In Thomas Kuhn’s The Structure of Scientific Revolutions (p. 146) he criticizes Popper’s falsification criteria as follows:

Nevertheless, anomalous experiences may not be identified with falsifying ones. Indeed, I doubt that the latter exist. … If any and every failure to fit were ground for theory rejection, all theories ought to be rejected at all times. (emphasis mine)

The famous Duhem-Quine Thesis says that all empirical hypotheses require one or more background assumptions. So it would seem that Kuhn is technically correct: it is impossible to falsify or refute a theory with a single (or any number) of anomalous observations because there is no way–using observations alone—to know if the problem was in the theory you are testing itself or one of the often implicit background assumptions.

This problem has long been seen as a disproof of Popper’s epistemology. How can we square Popper’s epistemology with Kuhn’s observation?

David Deutsch’s Solution to Kuhn’s Challenge

David Deutsch’s books have generated a lot of interest in Popper in recent years. Deutsch offers a possible way to deal with the Duhem-Quine thesis. In his paper, “The Logic of Experimental Tests…” he offers one possible solution to Kuhn’s criticism.

…in the absence of a good rival explanation, an explanatory theory cannot be refuted by experiment: at most it can be made problematic. If only one good explanation is known, and an experimental result makes it problematic, that can motivate a research programme to replace it (or to replace some other theory). But so can a theoretical problem, a philosophical problem, a hunch, a wish – anything. (p. 8, emphasis mine.)

But in any case, the existence of a problem with a theory has little import besides, as I said, informing research programmes – unless both the new and the old explicanda are well explained by a rival theory. In that case the problem becomes grounds for considering the problematic theory tentatively refuted.” (p. 10, emphasis mine)

So in Deutsch’s view, a theory is not refuted with a single (or any number) of observations. Rather it is refuted by a combination of anomalous observations plus an improved theory that explains under what conditions the old theory makes correct predictions. Or to put it another way: a refutation requires anomalous observations plus a better explanation. Until you have that improved explanation, you really only have a ‘problem’ rather than a ‘refutation.’ (A full discussion of Deutsch’s epistemology is available here.)

Popper Seems to Disagree

But if this is correct, then certain passages of Karl Popper feel problematic. For example, Popper said:

In so far as scientific statements refer to the world of experience, they must be refutable; and, in so far as they are irrefutable, they do not refer to the world of experience. (The Open Society and Its Enemies, Vol 2, p. 13)

If Deutsch is correct, then this statement from Popper is technically a false statement. It is not the case that scientific statements only refer to the world of experience if it is refutable because to be ‘refutable’ under the Deutsch concept of refutation one must also have an improved explanation as well as the anomalous observations. It would make very little sense for Popper to claim that for Newton’s theory to be considered a ‘scientific statement’ one must first have Einstein’s General Relativity.

Here Popper is even more specific:

…can the assumption of the truth of test statements justify either the claim that a universal theory is true or the claim that it is false? To this problem, my answer is positive: Yes, the assumption of the truth of test statements sometimes allows us to justify the claim that an explanatory universal theory is false. (Objective Knowledge, p. 7)

Again, if Popper had in mind the idea that a refutation required a second explanation than this seems like a technically incorrect statement because you can’t claim a theory to be false via the truth of a test statement alone due to the Duhem-Quine thesis. That would require an additional explanation as well.

And here Popper just comes right out and states that a single basic statement (i.e. observation) from a test can falsify a theory:

As to falsification, special rules must be introduced which will determine under what conditions a system is to be regarded as falsified. We say that a theory is falsified only if we have accepted basic statements which contradicts it.  (The Logic of Scientific Discovery, P. 66)

Nowhere in this passage does Popper say we need a second explanation to falsify a theory. To Popper testability, refutability, and falsifiability were synonyms: “Testability is, therefore, the same as refutability or falsifiability.” (Conjecture and Refutation, p. 266) So Popper clearly intended that we could falsify or refute a theory via testing alone, long before we come up with an alternative theory. So, something doesn’t square between Popper’s statements and Deutsch’s explanation.

Why People Dismiss Popper

To many of you that already know Popper well, the problem I’m outlining may seem more like a pseudo-problem because you already know how to resolve it. But I want to make the case that this is actually one of the main reasons why people reject Popper’s epistemology: because they can see that it is not possible to refute or falsify a theory with by observation alone even though Popper at times seems like he’s claiming you can.

Consider this video by Victor Gijsbers of Leiden University in which he claims to demonstrate that Karl Popper’s theory is incorrect. His argument is that Popper’s epistemology consists of “critically testing theories and showing there are false…” Gijsbers goes on to say “an observation that shows that a certain theory is wrong is called a falsification of that theory” and that therefore, in Popper’s view, “the only conclusion that scientists ever draw is a conclusion the theory is wrong.”

I suspect many Critical Rationalists would agree this is a decent summary of Popper’s epistemology.

But then Gijsber goes on to utilize a form of the Duhem-Quine thesis to show that this is logically impossible. “Popper was… wrong about the logic of falsification,” he says.

Gijsber uses an example of “’If all frogs are dead after a week in a the freezer’ is true, then this frog [left in the freezer] will be dead.” He points out that should we find the frog alive a week later we don’t really know the original theory to be false because we are making a number of additional implicit assumptions as part of our background knowledge. This might include that the freezer has not broken down or that some animal lover didn’t save the frog.

In fact, there are an infinite number of tacit assumptions that make up our background knowledge. A so-called ‘falsifying’ observation could easily be due to any of those infinity of assumptions rather than because the theory being tested is false. Therefore, Gijsber—just like Kuhn—concludes that there really is no such thing as a falsifying observation. Thus, Gijsber concludes, this is why most “…philosophers of science nowadays are almost unanimous in concluding that Popper was in fact wrong.”

Danny Frederick’s Solution To Kuhn’s Challenge

Like Gijsber, I was long troubled by Popper’s words. They left me with the impression that Popper thought a single anomalous observation can falsify a specific individual theory even without a second explanation. I knew this could not be true due to the Duhem-Quine thesis. For many years I just assumed Popper had misunderstood this aspect of his own theory.

It was in a conversation with Popper scholar Danny Frederick on social media that I finally understood the misunderstanding was my own. I told Frederick about how David Deutsch claimed you could not falsify/refute a theory without a second explanation and how, prior to having that second explanation, the anomalous observation was merely a problem to be solved but didn’t refute the existing theory. Frederick told me that was incorrect and that you could falsify/refute a theory with a single observation just as Popper claimed.

An argument ensued and as Frederick explained his view to me I came to realize something important that forever changed how I read Karl Popper.

I captured Frederick’s view in this blog post. Frederick’s view was that a ‘refutation’ was a single observation but that it ‘falsified’ not the theory you wanted to test but the combination of the theory plus the background assumptions

Frederick emphasized to me that due to the Ceteris-Paribus clause (meaning literally “with other conditions remaining the same”) the list of background assumptions was actually effectively finite at any given moment so a refutation of the combination of the theory plus the background knowledge was sufficient to allow us to come up with a testable conjecture as to what part of the theoretical system (i.e. the theory itself or the background knowledge) had caused the problem. If we performed the test and did not find the source of the problem, we could then conjecture that the problem was some other part of the theoretical system. So there was nothing stopping us from finding the source of the problem via a conjecture and testing process. It is literally problems and conjectured solutions all the way down.

What made this even more interesting was that Frederick had his own term for when you get a second theory and use it to dismiss the previous best theory. He referred to this as ‘rejection’ of the old theory rather than (as Deutsch had it) ‘refutation’ of the old theory.

A quick comparison between Frederick’s and Deutsch’s interpretations of Popper is below:

Frederick’s InterpretationDeutsch’s Interpretation
When you discover that an experiment doesn’t come out as a theory predicts, this is a ‘refutation‘ though it is understood not to be a refutation of the theory proper but a refutation of a combination of the theory and the background knowledge.When you discover that an experiment doesn’t come out as a theory predicts, this is not a ‘refutation’ of the theory but is instead merely a ‘problem‘ for the theory because you don’t yet know if it’s a problem with the theory proper or the background knowledge.
When you have a new rival theory that explains the above-mentioned refutation you can now tentatively ‘reject‘ the old theory.When you have a new rival theory that explains the above-mentioned ‘problem’ you can now tentatively ‘refute‘ the old theory.

Multiple Meanings for the Word ‘Refutation’

The above table shows that Deutsch’s and Frederick’s views are conceptually identical. However, they use different language to express themselves. Most importantly, each uses the word ‘refutation’ in a different way, though both are legitimate understandings of the word ‘refutation.’

So the word ‘refutation’ has more than one legitimate meaning! To David Deutsch one ‘refuted’ a theory by having problematic observations plus a new theory. To Danny Frederick one ‘refuted’ a ‘theoretical system’ with a single observation but by that he meant one refuted the combination of the theory plus the background knowledge.

Note that Frederick’s formulation directly addresses the criticism of Kuhn and Gijsber. To use the frog example, if the frog is still alive after a week in the freezer, the first thing you will probably check is if the freezer is broken (i.e. that your background knowledge was the source of the problem rather than your theory.) So the observation ‘the frog is still alive’ is a valid refutation of the combination of the whole theoretical system and this is all we actually need to move us to the next step: conjecturing how to solve the problem.

Popper’s Own Use of the Terms “Refutation” and “Falsification”

Now that I knew that Deutsch and Frederick understood the word ‘refutation’ differently, I realized that the supposedly problematic quotes from Popper (above) were now rendered unproblematic. But did this mean Popper intentionally used the word ‘refutation’ to mean a refutation of both the theory plus the background knowledge? The above quotes aren’t clear on their own. But now that I knew what to look for, it was possible to seek if Popper did understand a ‘refutation’ as such a combination.

For example, in Realism and the Aim of Science Popper makes a particular clear explanation:

[A more serious objection to my epistemology is] closely connected with the problem of context, and the fact that my criterion of demarcation applies to systems of theories rather than to statements out of context. This objection may be put as follows. No single hypothesis… is falsifiable, because every refutation of a conclusion may hit any single premise of the set of all premises used in deriving the refuted conclusion. The attribution of the falsity to some particular hypothesis that belongs to this set of premises is therefore risky… (REALISM AND THE AIM OF SCIENCE, P. 187. emphasis mine)

So Popper had always understood a ‘refutation’ to be a refutation of a combination of the theory plus the background knowledge. However, Popper goes on to explain why he sees this as unproblematic and offers the following reasons.

First, we usually choose to vary the background knowledge intentionally to reduce the chances that is the source of the problem:

…we consciously test… a certain chosen hypothesis treating the rest of the theories involved in the test as more or less unproblematic—as a kind of ‘background knowledge.’ This background knowledge is usually varied by us during the tests, which tends to neutralize mistakes that might be involved in it. (Realism and the Aim of Science, p. 188)

Secondly, the most important tests we do are crucial tests that we do between two competing theories.

…we always try to discover how we might arrange for crucial tests between the new hypothesis under investigation—the one we are trying to test—and some others. This is a consequence of the fact that our tests are attempted refutations; that they are designed—designed in the light of some competing hypothesis—with the aim of refuting, if possible, the theory which we wish to test. And we always try, in a crucial test, to make the background knowledge play exactly the same part—so far as this is possible—with respect to each of the hypotheses between which we try to force a decision by the crucial test. (Realism and the Aim of Science, p. 188)

Note how this second explanation is more or less the Deutsch interpretation of ‘refutation.’ So Popper really had both the Frederick and the Deutsch interpretation in mind, depending on context.

Misreading Popper

So it turned out that I had been misreading Popper for years. I had in mind that a ‘refutation’ was specifically a refutation an individual theory. This was why I had struggled to understand the passages from Popper quoted above.

But I was not alone in misreading Popper in this way! In fact, I’m in rather good company since Kuhn, Gijsber, and even the very Popperian Deutsch all happened to misread Popper in the same way.

What Does the Word “Refutation” Normally Mean?

In retrospect, this is an understandable problem. Try looking up the terms ‘refutation’ and ‘falsification’ and you may find definitions like this: (See this blog post for discussion)

Refutation: The action of proving a statement or theory to be wrong or false. (link)

Falsification: The action of falsifying information or a theory. (link)

Note how these dictionary definitions are specifically about refuting or falsifying an individual theory, not a combination of the theory plus the background knowledge. So the word ‘refutation’ does not, for the most part, carry with it a connotation of refuting a combination of a theory plus the background knowledge. So not surprisingly, to most people when you ‘refute’ something it’s the theory itself you refuted.

Given this is how the word is normally used, it is not surprising that Deutsch’s interpretation of that word required a second theory as only then can you really say that you refuted the theory itself, not in a combination with the background knowledge.

Yet it is impossible to read some passages of Popper trying to conceptualize ‘refutation’ in the Deutsch interpretation, thus leading people to think that Popper was mistaken.

Why Did Popper speak of ‘Refuting Theories’?

So if Popper saw a ‘refutation’ (or ‘falsification’) as applying to a combination of the theory plus the background knowledge, then why did he make statements like this?

…if the conclusions have been falsified, then their falsifications also falsifies the theory from which they were logically deduced… (LoSD, p. 10)

Nathan Oseroff in his paper “Addressing Three Popular Philosophic Myths about Karl Popper’s Demarcation Criteria” noted this problem of how Popper often comes across like he believes you can refute an individual theory (not the full theoretical system) with a single observation:

Reading Popper’s The Logic of Scientific Discovery would be sufficient for many people to arrive at the conclusion that falsifiability applies only to individual theories. (p. 6)

But Oseroff suggests that we should understand Popper as using the word ‘theory’ in situations like this as a shorthand for ‘a system of theories’ (i.e. the theory we’re interested in plus the background knowledge).

…any use of the word ‘theory’ in relation to [Popper’s] falsifiability criterion [should] be understood as an elliptical expression for ‘theoretical system’… (p. 10)

Making Popper More Clear

This creates an interesting problem: Popper often used the word ‘refutation’ and ‘falsification’ to refer to a combination of a theory plus the background knowledge but many people struggle to understand those terms in that way because it’s a bit unnatural. So misreading Popper is likely. Is there a way to restate Popper’s epistemology in a clearer way by not using those terms?

This is where the idea of “Popper Without Refutation” comes from. It’s not an attempt to in any way change Popper’s epistemology but to clarify it using terms that are clearer to others.

The different use of terms between Frederick and Deutsch are suggestive of how to go about this. Imagine that you used Deutsch’s term ‘problem’ (or better yet ‘counterexample’) to refer to an anomalous observation that ‘refutes’ a combination of the theory plus the background knowledge and Frederick’s term ‘rejection’ to refer to tentatively deciding to drop one theory in favor of another due to a crucial test between competing theories. You’d not need the word ‘refutation’ or ‘falsification’ at all yet you’d still be expressing an identical concept. Would this make Popper’s epistemology clearer to layman and philosophers that currently dismiss Popper due to this misunderstanding?

A Revised Popperian Lingo

This paper has concentrated on the words ‘refutation’ and ‘falsification’ which I believe tend to mislead people and lead to misunderstandings of Popper’s real intent. But eliminating those terms requires minor changes to other terms. As discussed in this blog post, the following table is my suggested list of changes:

Popper’s Old TermSuggested New TermDefinition
RefutationCounterexample / Anomalous ObservationA basic statement that contradicts the whole theoretical system (i.e. theory plus background knowledge.)
VerificationPositive Instance or OutcomeA basic statement that does not contradict the predictions of a theory because it matches the predictions made by the theory. A positive outcome really applies more to a test than a theory itself. See this blog post for a discussion on the often-misunderstood asymmetry between counterexamples and positive outcomes.
CriticismProblemTo Popper, a ‘refutation’ was specifically empirical. By comparison, a criticism might be a refutation or it might be some other form of criticism that might be non-empirical and thus can apply to metaphysical theories. I will more or less use criticism and problem interchangeably. So for me, a counterexample is a problem but not all problems are counterexamples.
 N/ARejection / AcceptanceWhen an individual scientist decides that they have found done a successful crucial test between an old and new theory and found that the old theory did not survive the test. This may require repeatability to be sure there was no mistake in the test. A synonym is ‘Acceptance’ of the theory that wasn’t ‘Rejected.’ Acceptance and Rejection are therefore logically identical. Popper downplays the importance of this event (See Myth of the Framework, p. 102-103), so Popper and has no equivalent for this term.
ConjectureConjecture Solution or Conjectured HypothesisI’m not really changing this one. But because Popper used the phrase “Conjecture and Refutation” as a shorthand for his epistemology we need a new shorthand that doesn’t use the term ‘refutation.’ “Conjectures and Counterexamples” or “Conjectures and Problems” does not sound right. So the new shorthand must now be “Problems and Conjectured Solutions.” As a bonus, this fixes the temporal ordering! Because according to Popper we always start with problems.

Conclusion

Karl Popper’s epistemology has not caught on strongly even among scientists much less philosophers. There is likely a variety of reasons for this. But one of the lesser reasons is that Karl Popper chose words that accidentally convey the wrong idea to the average reader. “Refutation” and “Falsification” as words too strongly convey to many the idea that we’re talking specifically about refutation of an individual theory rather than a full theoretical system. This tends to lead people like Kuhn and Gijsber and others to believe they can easily ‘refute’ Popper’s epistemology with a counterexample (such as the frog example) when really they are refuting a misunderstanding in their background knowledge.

In this paper I have suggested that Critical Rationalist would benefit from rewording Popper’s epistemology into terms that make Popper’s real intent clearer, particularly by dropping use of the terms ‘refutation’ and ‘falsification.’

I doubt this proposal is likely to catch on widely within Critical Rationalist circles as those terms have long since become integral to how Critical Rationalists express their philosophy. But it may be worthwhile to learn to use different language when speaking to non-Critical Rationalist.

Further Reading on the Problems of Understanding Popper’s Concept of Refutation

  1. A Summary of Deutsch’s Epistemology
  2. The Problems of Refutation
  3. Popper Explains The Asymmetry Between Refutation and Verification
  4. Do Deutsch and Popper Disagree Over Refutation?
  5. There is Nothing Wrong with the Language of Support
  6. Are Refutations and Verification Really Symmetrical Within A Theory Comparison?
  7. Demarcation: What Does it Mean to Be Empirical?
  8. But What If You Verify a Theory That Can Only Be Verified?
  9. The Two (or More) Kinds of Refutation
  10. How to Make Popper’s Epistemology More Clear

One Reply to “Popper Without “Refutation””

Leave a Reply

Your email address will not be published. Required fields are marked *