The Logic of Falsification

Blog post by Jeffrey Thayne on October 27, 2008

Jeffrey Thayne

In my previous post, I explained how experimentation is essentially a kind of logical argument. Based upon this assumption, I demonstrated how experimentation alone could never indubitably prove a hypothesis true, because whenever we conclude that our theory is true on the basis of observation, we have committed a logical fallacy known as affirming the consequent.

A philosopher of science known for pointing this out, Karl Popper, presented an alternative view of science, which is sometimes referred to as falsificationism. Rather than prove things true, he said, experiments should be designed to prove things false. The logical argument could be laid out like this:

If theory x is true, we will observe y.
We do not observe y.
Therefore, theory x is not true.

The logic of this syllogism is valid; there are no fallacies here. Thus, science would be useful for discarding false theories, one by one. From this point of view, there are no theories proven true; only theories that have yet to be falsified. This change of rhetoric avoids the logical mess of verification.

Popper’s point of view, however, is not without problems. Brent Slife and Richard Williams explain:

The strategy of falsification will not work unless we are sure that our test or experiment is the crucial test of the theory or hypothesis. No experiment will be a crucial test unless all possible variables (or limiting conditions) have been controlled or taken into account. There must be no other possible explanation for the failure of the experiment except the falsity of the hypothesis. This degree of control is, of course, impossible—practically and in principle. There are, in principle, an infinite number of things to be controlled in order to falsify any theory or hypothesis. Not all of them can be controlled, if only because there is no control over the particular point on the space-time continuum where any study is conducted—that is, each study is conducted as a particular place and time. Consequently, the effects of that particular context can never be controlled experimentally. [This space-time challenge is negated if you assume that time and space are homogenous, but that is an assumption, not an indubitable or demonstrated truth]1

In other words, “the methods of empirical science cannot falsify theories or hypotheses. This has been well recognized among philosophers of science (e.g., Lakatos, 1970).”1

As I explained in my previous post, my purpose here is not to dismiss science. I am not claiming that science is untrustworthy. My only claim is that the trust we have in science is not the inevitable result of the particular methods science uses. Many philosophers of science will agree that no method can be the royal road to truth; in other words, that there is no systematic way to obtain indubitable knowledge. “Whatever scientific methods may be good for,” Slife and Williams explain, “they cannot be used to verify theories, in the sense of affirming the consequent, or falsify theories, in the sense of negating the consequent.”1

This does not mean that science should have no persuasive power; I think we ought to be persuaded by well-performed, systematic, and thoughtful scientific experiments. Science is just that: a useful means of persuading people to understand the world in a particular way. I just don’t believe that science represents the progressive march towards truth that many people interpret it to be. Science may certainly unify us into a collective worldview, and persuade us to embrace a particular point of view. It isn’t, however, the road to certain truth.

In my next post, I will address why logic and reason are persuasive to us. I will also discuss a new way of understanding science that doesn’t involve the logical problems discussed in these posts.

1. Brent Slife and Richard Williams, “Science and Human Behavior”, in What’s Behind the Research? Discovering Hidden Assumptions in the Behavioral Sciences (Thousand Oaks, CA: SAGE Publications, 1995), pp. 167–204.

comments so far
  1. The classic problem with falsification is that one has to know something positive to be able to make a falsification. So falsification depends upon something known without falsification. The second famous problem is that any claim of falsification can usually be objected to by manipulating a system. Find an apple falling up and are you apt to say gravity was falsified or that there’s something else unknown going on?

  2. Exactly! Thanks for the comment!

    In order to falsify a theory, you have to rule out ALL other options by controlling ALL competing variables. Because that is not possible, falsification becomes impossible. You may be able to present a compelling, even convincing case that a theory is false, but never an indubitable one.

  3. And once it gets down to convincing cases falsification really has no more strength than verification. It becomes a more practical question whether one verifies or falsifies a theory. The important thing that Popper got right (despite most being wrong) was that there wasn’t some absolute standard of verification. Put an other way science is intrinsically a fallible process.

    I don’t think this ends up entailing it is merely about persuasion. For two reasons. For one science isn’t that persuasive to many Americans. Second there clearly is something about science that goes beyond mere persuasion which any sophist could do. That element of “surprise” is I think very key.

  4. Interesting thoughts. Would you explain what you mean by the “element of surprise”?

    I would agree that many Americans aren’t all that persuaded by science. I oftentimes count myself as one of them.

  5. I find your ideas about science and logic educating. I have not studied formalized logic very much, but I have thought that the limits of empirical science deserve examination. Which doesn’t mean that empirical science isn’t be the best way to study natural phenomena as far as we can tell. It’s just the conclusions we can draw from what we observe.

    But how far does the principle from quantum physics of the very act of observation affecting the observed phenomenon go? And how can we observe it (sorry, veered offtopic)?

  6. PAUL!? Paul Feyerabend, is that you!? Ha Ha, just kidding!

    Well done, Jeffrey. I wholeheartedly agree. What do you think about Popper’s opposition to induction? This is the other problem with his theory. He brushed aside Hume’s problem of induction by saying that his deductive model of science a’la falsification and corroboration did not involve induction. There is much more to say about it but I won’t waste space here. Let me just say that I think he was very much mistaken to think that he was avoiding induction.

    Lastly, while Popper’s falsification via the hypothetico-deductive approach is held up as the ideal today, I am sure you would agree that Thomas Kuhn presented a more realistic approach to understanding scientific progress.

  7. Velska, how to interpret the act of observation in QM is very much a controversial issue. So using it as a premise for further reason is deeply problematic. (i.e. you’ll find little agreement about what it all means)

    “LDSPhilosopher” by the element of surprise I mean that the unexpected tends to confirm science in a way general sophistry or persuasion doesn’t. I can persuade you of many things however reality has a way of interjecting that undermines your persuasion. It is that acting of reality – especially in unexpected ways – that is the great key to the success and rigor of science. It is why it is so unlike many other processes and why I think applying the term “persuasion” is misleading unless seriously caveated.

    BTW – you guys might want to put some leading padding to the paragraphs in your comments.

  8. Lastly, while Popper’s falsification via the hypothetico-deductive approach is held up as the ideal today, I am sure you would agree that Thomas Kuhn presented a more realistic approach to understanding scientific progress.

    I think the general consensus is that both have serious problems and neither can offer a “realistic approach.”

  9. Jeff I just discovered your blog. Very nice! I love the discussion.

    Actually though I don’t think falsification carries the weight it once did and in fact if the last 50 years of philosophy of science have taught us anything the scientific method can’t be boiled down to, well, a method. As I’ve written on my blog: Science consists of–logic, testing, trail & error, creativity, memory, falsification, confirmation, influence of current theory and paradigms, apprenticeships, refining technique, discussion, argument, going back to the drawing board, peer review, imagination, doubt, belief, asking questions, publishing, challenging convention, dreaming . . . and yes doing experiments where possible.

    The greatest strength of science lies not in methodological procedures but in (1) its openness to criticism so that mistakes are uncovered, bad logic recognized, and problems identified; and also its openness allows others to repeat any claim. (2) Peer review is vital in fact checking, making sure the study has context and coherence with other things that have been found and when anomalies between studies are found science’s theories provide guidance on tests that will distinguish competing hypotheses. Theory than plays the role of providing context for claims and is part of Lakatosean research programs that provide intellectual centers of thought and groups of researchers who are working within well confirmed areas of science; (3) Sciences most powerful tool of reason is abduction, a kind of inference. Not deduction. (I have post at my blog called “Alien Abduction” in May you can look at to see an example.

    But science is powerful. It has it problems but there is nothing stronger for sorting out the truths of this world and it has been wildly successful. Wildly! I get worried when I see the public reducing it to a method and then pointing out the weaknesses of those methods. It’s the vibrant back and forth among scientists that make it was it is. Fallible? Very. But likely to get corrected sooner or later and there is not better alternative on the horizon.

    I’m not dissing your post on this by the way. I love what you are doing. It just gave a too simplistic view of science and I think science too often gets misconstrued to be something it is not anymore and these views than raise suspicions in members of the church about science that do not recognize its strength and vitality.

  10. Jeff, you know we’ve butted heads on this topic before, but I’ve thought of an interesting argument that I think holds a grain of salt to support your side (though I don’t think I’m the first to think of it):

    It may be that it is impossible to prove a law or principle from within the system that holds the law or principle in question. For example, using the scientific method to prove the veracity of the scientific method is impossible, and we were unable to prove the commutability or associative nature of addition until we got into number theory, a deeper, less abstract system that helped us prove principles of our numbering system.

    So until we find that deeper, more object-oriented, less abstract system which can prove the scientific method, it remains only a method (though admittedly a method that has proved convenient for a large majority of scientists and brought about a great many important innovations and ideas). The fact that no deeper “meta” system exists for the scientific method, as well as the fact that it would be so wide-reaching and universal that the human mind can’t possibly imagine it, makes a great many people uncomfortable.

  11. Clumpy,

    I see what you are saying. However, my claim is that the scientific method can’t prove theories true or false (even those that fall within the system/scope of science). That is a claim that is virtually unanimously accepted by contemporary philosophers of science. The method may certainly marshall evidence in favor or against a claim, but never conclusively.

    I’m not trying to disprove the scientific method; I’m only saying that we need to be more modest about our appeals to science. Appeals to the scientific method may be persuasive, but hardly conclusive. Surely even the most ardent scientist will recognize the need for a simple humility about our truth claims.

    In the end, I believe that there is no royal method to truth; I don’t believe truth is something we arrive at through a method.

  12. I don’t think that I was very clear. I just meant that the scientific method remains a tool to be used only as we find it effective, since we can’t prove its truth or falsehood but it seems to be a good system for experimentation.

    The scientific method can’t be precisely encoded the way number theory can, so it can’t be logically proven or refuted. (And that doesn’t even get into the question of whether “logic” is a valid method of proven the truth of something, as your “royal method to truth” statement says.)

  13. SteveP: I think science too often gets misconstrued to be something it is not anymore, and these views than raise suspicions in members of the Church about science that do not recognize its strength and vitality.

    I see what you mean, like if someone were to distrust a new medicine or something because it was “only” based on science. We owe a lot of our leaps and improvements to that active scientific discussion.

    I think, though, that when people point out science’s weaknesses, they’re often trying to help people understand and apply science more accurately. A person’s gonna use a hammer better when they understand that it isn’t good for driving screws. Understanding the hammer’s limitations means people will use it better.

    But you’re right: if all craftsmen only emphasized that hammers aren’t any good on screws, while forgetting to mention that they are good for driving nails, then fewer people would use hammers, and to their own loss. Yes, the craftsmen should point out how good hammers are on nails.

    There is not a better alternative on the horizon.

    In the true domain of science, sure. But there are several better alternatives to science in other domains.

    When people frequently refer to science in order to make, for example, large-scale societal changes of a moral nature, it’s clear that far too many people see science as the wonder-tool that not only pounds nails, but drives screws, cuts lumber, and peels potatoes as well. That’s when I think it’s healthy and useful to give frequent reminders of what the hammer is for, and what it is not for.

Leave a Reply