Falsifiability (Column 403)

With God’s help

Disclaimer: This post was translated from Hebrew using AI (ChatGPT 5 Thinking), so there may be inaccuracies or nuances lost. If something seems unclear, please refer to the Hebrew original or contact us for clarification.

In the previous column I dealt, among other things, with the raven paradox. There I mentioned Popper’s criterion of falsification and the philosophical debate about the confirmation of a scientific theory. In this column I wish to touch on falsification and on the scientific status of theories.

A first look at science and a scientific theory

Many people think that science deals with proving theories, or alternatively, that a scientific theory is a theory that has been proven. To examine this, we must define the very notion of “proof.” If one means proof in the sense of a valid logical–deductive derivation, it is clear that you will not find that in science. Science is tied to observations, whereas mathematics and logic are a priori domains. If a scientific theory were proven in the logical sense, there would be no need for observations, and science would effectively be part of mathematics. If that were the case, a scientific theory would also not add information about the world (because the analytic is empty; see Two Carts in the first gate). But the aim of science is to collect information about the world and organize it in a systematic theoretical framework—that is, to create theories. In this sense it cannot be analytic and empty, and therefore it also cannot be based on logical proofs.

Yet contrary to a common but opposite mistake, it is equally incorrect to say that science (unlike mathematics) is based only on observations. Science indeed uses observations, but the scientific theory is not built on observations alone. Observations pertain to a limited number of particular cases that we have observed, whereas the theory concerns a whole class of cases, most of which we have not observed. The theory is constructed as a generalization based on the specific observations we have made. The process of generalization takes us beyond what we directly saw in our observations. Observations are individual instances, and the theory deals with general laws; that is, the inference goes from particulars to the general, and therefore it cannot be deduction (which goes from the general to the particular). Thus, the generalization does not follow necessarily from the observations, and it contains a measure of speculative thinking.

What, then, is the standing of a scientific theory? Is it mere speculation? What role do the observations that underlie the matter play? One can even wonder about the difference between a theory based on observations and a theory I concocted in my fevered imagination. Philosophers of science will tell you that the difference lies in the existence of predictions. A scientific theory is supposed to yield predictions—that is, it should present the results of future experiments that can be performed. But this does not exhaust the scientific process, since even from an imaginary theory one can derive predictions (the chance they will work is probably smaller, but that is not a definition of scientificity).

Others will say the difference lies in the origin of the theory. A scientific theory is based on observations (even if it is not deductively derived from them) rather than on dreams at night. A theory whose origin is speculation rather than observations is not scientific. But this too is not a satisfactory answer. In principle, there can be a scientific theory that was conceived in a dream, and as long as it passes the empirical test (though the likelihood of that is smaller, as noted), we ought to adopt it. In the philosophy of science it is customary to distinguish between the “context of discovery” and the “context of justification.” The context of discovery concerns how the theory arose (was discovered), whereas the context of justification addresses how it is justified once it has already been formulated. The context of discovery carries no scientific significance. We have no problem with a theory that arose from Elijahic revelations, hallucinatory visions, or any other source. Our demand of a scientific theory is that it pass empirical justification tests (that it “works”). That is the theory’s context of justification. If it works, there is no reason to reject it merely because of its dubious origin.

Bottom line, we remain with the question: what is the difference between the way a scientific theory is created and a fabrication? More precisely: what defines the scientific nature of a theory at all? The way a theory is generated from observations is called by logicians induction or abduction (see on this in columns 379 and 399). This is a non-deductive inference technique, and yet we still regard it as rational rather than arbitrary. I have discussed this quite a bit in the past. Still, it is hard to rely on such a criterion (=everything formed by inductive inference from observations is scientific) as defining a theory’s scientificity. Here we are after a criterion for that.

Popper’s criterion of falsification

Because of these doubts, there are philosophers of science who see science as a kind of game and not a distinct discipline. Moreover, some hold that the scientific theory is actually a claim about us rather than about the world. It is not true that there is a law that any two masses attract one another (the law of gravitation). What is true is that, in our experience so far, every pair of masses we have seen has indeed done so. The generalization from here to all cases is unfounded, and therefore it should not be treated as a serious claim about the world itself. They simply do not place their trust in induction and abduction, for the reasons described thus far.

To move toward a sharper criterion of scientificity, Karl Popper (see, for example, in his well-known article here) proposed the criterion of falsification: a theory is scientific if and only if it is falsifiable. The basis for his view is a surprising asymmetry between proving a theory and refuting it. Take as an example the scientific theory: all ravens are black. You cannot prove such a theory (or any other scientific theory), since to prove it in the full logical sense we would have to observe all the ravens in the world. Even if we succeeded in doing that, and even if we somehow knew that we had indeed seen them all, in such a case the claim would cease to be a theory and would become a summary of a collection of observations. A theory, by definition, also tries to speak about things we have not yet observed (and to give predictions about them). A summary of what we have seen is not science (but at most a useful instrument for scientific generalization). By contrast, refuting this theory is very easy in principle: we must find a single raven that is not black. That does not mean we will indeed manage to see such a raven (if the theory is true, then certainly not), but the very possibility of formulating an experiment by which the theory can be refuted is the criterion for its being scientific.

Thus, Popper argues, a theory is scientific if it is falsifiable, i.e., if it offers a prediction for the outcome of an experiment that can be performed. If we carry out the experiment and get a different result—the theory has been refuted. If we get the expected result—the theory has not been refuted (but of course also not proven), and in that case it can still be used. It is important to emphasize that a theory need not actually undergo a falsification test to be deemed scientific. It must supply a prediction that can be tested in the lab if we choose to try to refute it. Hence a scientific theory is not necessarily a true theory. There are also false scientific theories—these are theories that provide predictions (and are therefore scientific) some of which have been refuted (and are therefore false). Even theories that have not yet been refuted may of course be refuted in the future. The very falsifiability, Popper claims, is the criterion of a theory’s scientificity.

What is not a scientific theory according to Popper?

This criterion is, of course, very minimal. It is interesting to examine what, according to it, is not a scientific theory. A moral theory, for example, is not scientific. There is no way to subject to a refuting experiment the theory that murder is morally forbidden (as distinct from the theory that murder is legally forbidden—you simply open the statute book). The same holds for claims about the existence of God, which obviously cannot be framed in a way that refutes them. Try to think of an experiment that would show that there is no God, or that there are no demons, or that a demon has three wings and not four, or that murder is a good deed. All these are claims that cannot be subjected to an empirical test, and therefore they are not scientific. The same applies to determining the number of ants in the world. That is presumably not a scientific theory, since there is no empirical way to check it.

There are subtler examples. For instance, the theory that the Torah was given by the Holy One, blessed be He, and therefore the prophecies that appear in it (or in the Prophets) come true. It is not plausible to regard all these as scientific theories, since it is hard to fix measures for the fulfillment of such a theory. Is one woman who eats the flesh of her children a fulfillment of a prophecy? That can happen under harsh conditions, and there is no need to posit prophecy based on supernal knowledge. The same holds for exile or return from exile. For our purposes, what matters is falsification: if this did not happen throughout all history, is that a refutation of these claims? Perhaps it will yet happen?

So too regarding the “prophecies” of the oracle at Delphi, who always took care to word them vaguely enough precisely so that they could not be refuted. The same holds for modern academic oracles (part of what, through no fault of their own, is called “prediction science,” or futurology) who, exactly like their ancient father in Delphi, are very careful to use wording vague enough to resist scientific examination (see on this in columns 88 and 298). You can hear from them “predictions” like: “In the coming decade there will be economic prosperity,” “The State of Israel is heading toward a severe crisis,” “Our economy is expected to undergo upheavals,” and the like. And I have not yet spoken about the quality of the predictions (that is, that even in the rare cases where there is a claim that can be tested, its predictive quality is no better than that of Paul the octopus). Here I have dealt only with the quality and structure of their predictive claims (namely, their vagueness). When science gives a forecast it is supposed to give a precise quantitative forecast (with estimates of error bounds), and thus it can be subjected to a refutation test. Oracles of various kinds are careful not to do this (and when they do, on rare occasions, their forecasts, to my impression, are at grass-height).

The law of small numbers

In column 38 (and several others) I discussed a phenomenon I called “the law of small numbers.” People latch onto specific examples and infer general conclusions from them. Most alternative medicine (after all, my grandmother was in a state where all the doctors despaired, and she only took drops X and the next day she was good as new) and mystical medicine (rabbi So-and-so gave a blessing, or Oren Zarif treated according to the well-known Kabbalistic method, and it always works. The proof is that my grandmother…) is built on this fallacy.

What is the problem with drawing conclusions from certain examples? Seemingly this is science’s way in every field. Beyond the distinction between reasonable and less reasonable generalizations, here I wish to focus on Popper’s principle of falsification. The question is how we test such claims, namely whether we subject them to refutation tests. To subject such determinations to a refutation test, they must be defined more precisely. Method X works at a certain level of statistical significance, and we must test this against a control group while maintaining double-blind and so forth. A single case says nothing, since the question is what happens in large numbers. As long as the determinations are not presented in a quantitative and measurable way, they are not scientific and are therefore suspect of charlatanry. The law of small numbers is an expression of failure to meet the criterion of falsification.

Put differently, one could say that these claims are scientific, since they can be subjected to refutation tests. The problem is that people do not actually do so, and in the rare cases where it is done, there is no real evidence for the effectiveness of these methods. Below I will explain that the problem stems precisely from the fact that these are scientific claims.

Junk science

It is a common refrain in my mouth that many academic fields do not deserve the title “science.” I am not speaking about fields such as gender studies, practical and experiential Kabbalah, or prediction, which do not deserve to be called “academy” at all. I mean fields like literature, psychology, political science, history, philosophy, and the like (most fields called the “humanities” and “social sciences”). These fields generally do not present claims that can be empirically examined, and therefore there are no claims there that can be refuted. For example, I do not know a literary scholar who can give me in advance a quantitative prediction of phenomena in some literary text, or supply a claim that can be refuted on the basis of his professional knowledge. Literary studies deals mainly with sorting and classifying various phenomena, diagnosing texts, and mapping poetic phenomena. That can of course be interesting, and sometimes even intelligent, but it is certainly not science.

The same holds for most work in psychology. Usually we are dealing with qualitative insights that are very hard to measure and verify, and in most cases they are also not formulated in quantitative language. Regarding psychoanalysis, Popper already pointed out its non-scientific and speculative nature. Again, this does not necessarily mean that psychology has no value (which, in my opinion, is also largely true). Here I only claim that it is not scientific. Nowadays there is a major effort to ground psychology scientifically—that is, to formulate theories that are falsifiable (in particular its seepage toward neuroscience points to this)—but to my impression this usually does not really succeed. One can quantify various phenomena and perhaps even offer predictions, but in most cases, to my impression, we are not dealing with claims that can truly be quantitatively tested and refuted.

Moreover, even if in theoretical (or descriptive) psychology you can find claims and studies that meet Popper’s criterion, their clinical application in therapeutic methods or analysis, in the vast majority of cases, is anything but scientific.[1] It is closer to an art than to a science. Personally, I am quite skeptical even about the successes of such treatments, and even more so about the discipline’s contribution to those successes (the law of small numbers is broadly applied there).[2] Very few of these techniques have subjected treatment outcomes to quantitative scientific tests. As noted, our concern here is only with scientificity and not with the question of whether it is wise, interesting, or even effective.

Confirmation

According to most philosophers of science, Popper offers too thin a criterion, if any, and therefore there has been much criticism of his criterion. Some have argued that it is not correct that if a theory is refuted it is false. There are powerful theories that explain many situations, and with respect to them, even if we find several empirical refutations, we will not immediately discard them but wait for some solution or correction of the theory (or for an experimental error to be discovered). But this is mainly a practical claim. On the principled level, it too concedes that a theory that has been refuted in a single experiment is false. Others have argued that Popper’s criterion is too broad. Under the heading “science” it includes claims that are very hard to regard as scientific. For example, the theory that hammering a nail into a sandal fastens the straps to the sole. It certainly offers predictions, and it is probably also true (and likely will not be refuted in the future), and yet it is hard to treat this as science. It is simply too trivial and very non-abstract. But this too is largely a technical claim, and Popper would likely say that this is indeed a scientific theory, even if a simple one.

Others have pointed out that Popper’s criterion is too minimalist, since according to him there is never a true or acceptable theory—there is only a theory that has not yet been refuted. Is it reasonable to rely on a theory merely because it has not yet been refuted?! There are quite a few theories that have not been refuted and yet no one considers taking them seriously. Bertrand Russell’s celestial teapot may illustrate this. From here the idea of confirmation (corroboration) of a theory naturally arises. Various philosophers have argued that each experiment in which a theory’s prediction is borne out strengthens (corroborates) it. That is, a theory that provides predictions and all of them are confirmed in experiment is a strong theory (and not merely a theory that has not yet been refuted, as Popper claimed). Moreover, the larger the number of predictions confirmed—the more corroborated the theory is.[3]

It is very likely that Popper too accepts this claim, and in my view even those I mentioned above who doubt the validity of scientific laws (and see them as claims about us rather than about the world) do not dispute it. I assume all these people have flown on airplanes and placed their trust in the aerodynamic theory and materials science on which the airplane is based, and did not fear that the plane would crash. If these theories are only claims about us or merely not yet refuted claims, it is hard to believe anyone would entrust their life to them. Therefore it seems to me that almost no one truly disputes the thesis of corroboration, including Popper. As I understand it, Popper was only seeking a minimal and logically sharp criterion for defining a theory’s scientificity. Corroboration cannot be quantified, and one cannot draw a line as to how corroborated a theory must be to count as scientific. Moreover, corroboration concerns the truth of the theory, not its scientificity (above we saw that these are two different things). In my understanding, Popper merely sought to exclude corroboration from the bounds of the philosophy of science, but that does not mean he did not accept it in his everyday thinking. In his view too, a theory that passes many tests and whose predictions are confirmed is a strong and acceptable theory. But in his view this itself is a common-sense assumption, not a scientific claim. If so, it appears to be a merely theoretical, semantic dispute.

Either way, Popper’s criterion for a theory’s scientificity is accepted as a minimum criterion by almost all philosophers of science. Accordingly, we can exclude from the scientific sphere theories without predictions like those mentioned above. Claims about God, demons, the soul, moral claims, or non-univocal claims in various fields (like vague forecasts), and the like. This is not because these claims are false, but because they do not provide predictions that can be subjected to an empirical refutation test.

The deeper significance of falsifiability

We have seen that falsifiability is a formal criterion for a theory’s scientificity. But why is this important at all? Why does it matter whether a given theory is scientific or not? The important, relevant question is whether it is true or not, and the question of scientificity would seem merely a semantic definition. It appears the relevance of this discussion stems from the weight science carries in our thinking. A scientific claim is perceived as stronger than a non-scientific one. In other words, a claim that is falsifiable appears stronger than a claim that is not falsifiable. This is not only a semantic question of what deserves to be called science, but one of the claim’s degree of force and weight.

What is the reason for this? For a claim that is not falsifiable, the one who uttered it assumed no risk, and therefore did not have to weigh his words carefully. Note that this is not a flaw in the claim itself but in the claimant. Therefore, when I examine a claim on its own merits, the fact that it is not falsifiable is not necessarily a flaw. For example, vague forecasts given by an oracle, ancient or modern, are weak because he assumes no risk. He can offer a forecast, and if it succeeds he garners renown. There is no chance it will be refuted, and so this is unearned fame. He assumes no risk; therefore it is hard to ascribe significant weight to his words. Likewise, a mystical or alternative healing technique assumes no risk, since we will never be able to determine that it failed. One can always blame something else. Therefore its success is not worth much either.

But here we already see that beyond the risk assumed by the claimant, there is something problematic in the claim itself. True, this is not the strength or weakness of the claim but of the success of its prediction. That is, the claim that taking an alternative potion or performing mystical action X brings healing is not weaker because it is not falsifiable. We saw that the one who offers it is less credible (that is, we need not assume he examined carefully), but this has no bearing on the credibility of the claim itself. I now wish to add that if we subjected it to a refutation test and it passed, the success does not strengthen the thesis. Where the thesis cannot fail, there too its success is not worth much; that is, it does not strengthen it.

In sum, three things can be said about non-falsifiable claims:

• The one who presents them is less credible.
• If they succeed, that success carries little weight.
• But the claim itself is not necessarily less true because of this.

Does non-falsifiability contradict the possibility of success?

In the analysis above I assumed the possibility that a non-falsifiable claim can succeed. This itself seems odd at first glance. After all, if a claim can succeed, then its failure would be a refutation. The success of a claim means that its prediction is confirmed in an experiment. If so, to say that a theory can succeed implies that it has predictions. But if it has predictions, then it can also fail (when the prediction is not confirmed in the experiment). How, then, can there be a non-falsifiable claim that can succeed?

It turns out that Popper’s criterion of falsification is not as sharp as it first appears. Take the prophecies in Scripture, for example. Suppose one of the prophecies was fulfilled in a striking way. This is a success of the biblical prediction, i.e., support for the theory that the author of Scripture has abilities and supernal knowledge (or control) of the future. On the other hand, I have already explained that the failure of such a prophecy is not necessarily a refutation of that thesis. One can always argue that it has not yet happened but will happen in the future, or that it did not happen because of some side reason, etc. If mothers eating their children under siege comes to pass, that is an impressive fulfillment of a biblical prophecy. But if it does not happen, one can always suppose it will yet happen, or that it was merely a threat that need not materialize, or that it is a metaphor for severe suffering we will undergo, and so on. Does it follow that because of all these excuses for cases of failure, the fulfillment of the prophecy (i.e., success) necessarily carries no weight? In my opinion, if such a prophecy is fulfilled it is impressive, even though if it is not fulfilled I do not necessarily give up on the “theory” (trust in Scripture). One can say that such a fulfillment corroborates the claim (the credibility of Scripture), even though it is not falsifiable. This is already a conceptual distinction between corroboration and refutation, since the common view is that non-refutation is necessarily corroboration, and non-corroboration is refutation.[4]

I have already brought in the past R. Elchanan Wasserman’s example from the Talmud which says: “He who recites Havdalah over a cup of wine will have male children.” This is difficult, since there are those who recite Havdalah over a cup and yet have no male children (the Chazon Ish had no children at all, and it is very likely he was meticulous in reciting Havdalah over wine). R. Elchanan explains that there can always be other considerations that prevent the prediction from materializing. For example, there is also a Talmudic statement that “because of vows, a man’s children die.” So if a person is meticulous about Havdalah over wine but not meticulous about vows, what will happen to his children? It is indeed possible he will have none. Likewise with other side considerations that can intervene in the theory’s prediction (for example, if the man’s wife is infertile, then even if he recites Havdalah over a cup, he will not have children). In my article on “okimtas” I discussed this phenomenon at length and showed that it exists in scientific domains as well. Thus, there can be side causes that serve as ad hoc excuses for the failure of a theory’s prediction.

From here another conclusion relevant to us arises: as has often been noted, almost no theory is scientific in Popper’s unequivocal sense. For almost any scientific claim that fails in an experiment, ad hoc explanations can be given to save it from refutation (there was an error in the experiment, the temperature was unsuitable, the instrument was not clean, and the like). That is, even falsifiability is not as sharp a criterion of scientificity as Popper thought.

Non-scientific domains

There are domains for which there is, by their very nature, no scientific way to treat them. Metaphysics, literature, or ethics are good examples. In particular, the question of God’s existence is an interesting test case. Because belief in God is not falsifiable and is not scientific, is it therefore a weak or implausible claim? True, there is no possibility of obtaining for it either corroboration or refutation on the direct empirical level, but still there is no necessity to dismiss it. To this end we must note several points.

First, the way we arrive at belief is very similar to scientific generalization. We start from observational facts (that there is a complex world, and perhaps also that a complex thing is not created by chance[5]) and by way of abduction arrive at the theory that there is some cause that created the world. The inference is very similar to scientific generalizations, and if we trust those there is no reason not to trust this. Of course, this does not make God’s existence a scientific claim. It is not—if only because it is not falsifiable.

Second, we are dealing with a domain that is, in principle, inaccessible to empirical observations, and therefore the absence of predictions is not necessarily a weakness of the claim. In the terms used earlier, there is certainly no suspicion toward someone who presents the claim about God’s existence as if he is evading empirical testing. This is a domain that, by its nature, is not accessible to empirical examination. True, it remains the case that it is easy for an interested party to raise such a claim arbitrarily, for it cannot be tested.

Third, we saw above that a non-falsifiable claim is not necessarily weak because of that. In this case, there exist very good arguments supporting it (as noted, similar to scientific inferences; see my book The First Being), and therefore there is no reason to reject it just because it is not scientific. Rejecting it on the grounds that it is not falsifiable implicitly assumes that science is our only tool for gaining insights—an entirely unfounded assumption (certainly not empirically supported) that unjustly dooms all those domains.

It is important to note, however, that the non-falsifiability of such a claim does, of course, prevent us from testing and corroborating it, and that is indeed a special weakness of it. But, as noted, that is the nature of the domain. Think of a person who tells me he saw something. His claim is not falsifiable, since I cannot subject it to an empirical test. I can believe him or not. Because that is the nature of the claim, I must decide by non-scientific tools whether he is credible. And it is still not reasonable on that account to determine that he is not credible. Similarly, in deciding whether God exists, we have no option of using scientific tools. That does not mean there are no truths here, nor that such truths are inaccessible to us.

The conclusion is that in non-scientific domains there is no inherent deficiency in non-falsifiable claims, neither in the claimant nor in the claim itself. Such a claim should be examined by the tools relevant to the domain in question. I note that from Popper’s criterion there is no significance to the fact that such a claim cannot be refuted, but if we take corroboration procedures into account, then we must note that we have no way to corroborate such claims. That is some weakness of theirs, even if it stems from the nature of the domain in question.

Pseudo-scientific domains

There are quite a few domains where the excuse that the domain is not accessible to science is mere evasion. Such claims arise with respect to the fulfillment of vague forecasts, the success of prayers, alternative medicine, and the like. More than once people who have heard my criticism of these things have told me that I should not apply scientific thinking to them because they operate on other planes and with other tools.

The problem is that the fact that these domains do not use scientific tools is only by their choice. They can certainly do so, since their results are just as accessible to empirical examination as any scientific claim. The outcomes of mystical or alternative medicine can be tested via treatment and control groups, while maintaining double-blind procedures and standard scientific rules, just like conventional medicine. The fact that this is usually not done (and when it is done, it usually turns out there is no evidence of their effectiveness) arouses heavy suspicion about them—and rightly so. The same goes for forecasts and prophecies. Presenting forecasts vaguely prevents us from testing them, whereas if there were anything to them I would expect a more unambiguous presentation that can be empirically examined. Therefore, here too it appears that we are dealing with evasion, and these claims are highly suspect.

These domains, unlike metaphysics or ethics, are in principle scientific. The claims raised there are scientific, and precisely because of that, the fact that they cannot be subjected to critical testing (or that this is not done in practice) arouses heavy suspicion about them. Suspicion, of course, does not mandate outright rejection. As we saw regarding belief in God, it depends on the other evidence I have for these claims themselves. If someone raises a good claim in a suspicious way, I will examine the claim on its merits and form my view of it accordingly (context of justification rather than discovery). If I had a source that grounds such claims, I would be prepared to accept them despite the suspicion (as, for example, with the prophecies in the Torah). As explained above, such suspicion concerns the person presenting the claim rather than the claim itself.

Tautologies

There is another kind of non-falsifiable claim for which it is, in fact, clear that we ought to adopt it. I mean mathematical or logical claims, i.e., tautologies. I will not elaborate here because I have already done so in the past (see on this in columns 50 and 318), and I will suffice with presenting the matter briefly to complete the picture.

A logical argument is not falsifiable (either logically or practically; see on this distinction in columns 50 and 341), but that is precisely because it is necessarily true. For example, the statement 2+3=5 is not falsifiable. No experiment can refute it because it is necessarily true. The same goes for the claim about “survival of the fittest” in evolution. It is not falsifiable, because if some organism survived, it is necessarily fit (otherwise it wouldn’t have survived). But contrary to some critics of evolution, this statement is not a critique of the theory but merely a description and characterization of it. The fact that this is not a scientific claim does not mean we should not accept it, or that we should suspect it. On the contrary, it is not falsifiable because it is necessarily true by virtue of its very content.

The conclusion is that tautologies are indeed non-falsifiable claims, but this does not contradict the fact that they are (necessarily) true. Such non-scientific claims we are obliged to accept. This is another example that Popper’s criterion is a criterion of a claim’s scientificity but not necessarily of its truth or credibility.

[1] See, for example, the discussion here and here. However, in the pair of articles (here and here), Eyal Pozniak systematically examines the issue, and also distinguishes well between descriptive psychology and clinical psychology. He notes that on the clinical level, methods known as CBT do try to meet scientific criteria and are even supported by scientific studies that have quantitatively tested their effectiveness (see also in Wikipedia there, under “Research support”). He concludes that there is some effectiveness to psychodynamic therapy (and he greatly qualifies the meaning of this conclusion), but he focuses on comparisons between the methods and not on testing the objective effectiveness of psychology or on the discipline’s contribution to that effectiveness (e.g., comparison with conversations with coaches or with rabbis/Hasidic rebbes who do not use disciplinary tools). And even with respect to comparisons between psychological treatment methods, he himself mentions the phenomenon of “convergence to equality,” i.e., the fact that studies that try to test this quantitatively fail to find a difference between the different treatment methods. The issue of psychology’s scientificity has arisen here more than once, and I still plan to enter it in the future in a more focused and systematic way.

[2] In my estimation, even when there are successes (which are measurable and not placebo—rare, in my opinion), they result more from good intuition and experience and less from disciplinary knowledge.

[3] Incidentally, corroboration is related to what I discussed in the previous column. The claim here is that the more empirical tests a theory has passed, the greater the probability that it is true. This is not a strict entailment but a probabilistic (or plausibility-based) inference.

[4] Above I explained that, in my view, Popper too accepts the concept of corroboration; he simply does not grant it scientific status.

[5] One can debate whether this is an observational or an a priori claim.