Sometimes I want to talk about CR rather than Popper, to “de-Popperize” CR and give more credit to others but still Popper is the 800lb gorilla in the room. It is amusing to see the lengths that the philosophers have gone to avoid taking notice of him, like inventing the term “disconfirmation” so they don’t have to say “falsification”!
This post continues the “playing fields of Vienna theme” in a more positive manner to suggest how the rift between Popper and the other Austrians can be fixed. There are three aspects to this:
1. Ideas from Popper that have a critical or error elimination impact, where the positive effect is to economise efforts by indicate some things that don’t need to be done.
2. Ideas that support the things that need to be done, which the Austrians tend to be doing already.
3. Another type of error-elimination, pointing out the way some of the Austrians have contributed to the debacle of scholarship in philosophy by getting Popper wrong (samples are in the “playing fields” post).
Error elimination.
The Austrians, starting from Mises have sunk a lot of intellectual capital in supporting methodological dualism, to prove that positivism and empiricism (which they assume to be the appropriate methods for the natural sciences) will not work in economics. However in the light of Popper’s work we can see that positivism and empiricism do not work in the natural sciences. Popper’s alternative can be labelled “conjectural apriorism” to make the connection with the “fallible apriorism” which Barry Smith located at the heart of Carl Menger’s program in economics. That case is argued here in a paper that has been submitted for a collection on the Rothbardian contribution to the economics of von Mises which will appear in an Italian journal if the referees like it.
If that argument turns out to be robust then it means that Austrians such as Hans-Hermann Hoppe can stop defending strong versions of apriorism, and simply get on with their work in economics and political economy, developing theories that solve problems and stand up to criticism.
In this connection I will briefly mention a paper by Leeson and Boettke which did not cite Popper, so it can’t really be noted under the heading of the debacle of Popper scholarship but it did mention “falsification” in a rather strange way. The point of the article was to explain that von Mises had a more subtle and robust position on the methods of economics than most people realised, friends and critics alike. They drew on Barry Smith who is the most helpful guide on these matters but did not invoke Popper and point out that when you take on board Smith and Popper you end up at the same place, so you can indeed “put a fork in your philosophy” as Peter Lesson advised in a different context and get back to serious work in economics.
“For most of his career Mises found himself in a methodologically uncomfortable position. As a German-language economist, the discipline he was educated in was dominated by historicism. As a Viennese intellectual, Mises began to mature as a thinker within the philosophical culture of Wittgenstein and the Vienna Circle. When he published his first major statement concerning his methodological views (1933), logical positivism was starting to spread in economics. As Mises conceived it, logical logical positivism denied the existence of a priori knowledge and rejected all non-empirical forms of analysis (see Greaves 1974). According to this view, if economics is to progress as a science, indeed, if it is constitute a science at all, it must follow the methods of falsification employed by the physical sciences.”
The exponent of falsification who was cited at that point was Hutchison who achieved a degree of fame by invoking Popper as a corrective to certain lines of irrationalism such as Marxism and National Socialism that Hutchison perceived to be distoring the discussion of economics and public policy. However Popper’s ideas suffered much the same fate as those of von Mises, being more subtle and nuanced than most of his friends and all of his enemies realised. That would have been a useful point for Leeson and Boettke to make because others like Blaug have followed Hutchison to make use of rather blunt and unhelpful version of falsificationism as though the idea of testing was Popper’s only (and entirely unoriginal) contribution to the philosophy of science.
I have read the Leeson and Boettke paper several times and I am still wondering what is the take-home message that will help the working economist and how they can hope to advance a discussion of epistemology and methodology without referring to the 800lb gorilla, arguable the most important epistemologist and methodologist of recent times.
TO BE CONTINUED
Rafe,
My perception is that reticence about the term “falsification” and “falsificationism” is a result either of confusing logical issues with methodological ones, or of replacing methodological formulations entirely with logical ones. Popper’s critics do this constantly, because it is only by focusing on logical rather than methodological issues that they can get any traction with many of the old, standard criticisms. For instance, Larry Tapper in the Critical Cafe constantly repeats standard observations about falsifications being nothing more than inverted verifications despite the fact that he has been told repeatedly that falsificationism is a methodological prescription, not a logical one. These logical equivalences are completely irrelevant. Methodologically, attempting to falsify a theory is not the same as trying to verify it’s negation. The focus on logic instead of methodology leads Tapper and others like him to the fantasy that there is something called “Popperese” — a cultish, jargon-ridden language in which people are superstitiously enjoined to talk in an upside-down fashion so that they can pretend falsifications aren’t also verifications and so forth. Once the point about methodology is taken on board, however, these criticisms all fall to the ground. And of course, “naive falsificationism” is just a label for the most common manifestation of the failure to focus on methodology instead of logic. It is a view that Popper never held.
My take on the terminological issue, therefore, is that I wouldn’t let the critics have the word, i.e., I wouldn’t back away from the word “falsificationism”. In this I believe I am in agreement with David Miller. Terminology aside, I highly recommend all of his essays in _Out of Error_. The one of special relevance here is “Falsifiability: More Than a Convention?”
Regarding the term “apriorism” — the reason I don’t use that term is that it places the emphasis on origins, which in this case amounts to placing it on the subject, i.e., the individual knower. In what sense is knowledge said to be a-priori? It is said to be so from the point of view of the individual, the subject. Such knowledge is allegedly prior to sense experience. But ideally, the subject shouldn’t enter into the picture. We want the focus to be on knowledge in the objective sense.
Rafe,
One more note on the methodological asymmetry.
The reason that falsifying ‘All swans are white’ isn’t the same as verifying ‘There exists a non-white swan’ is that the methodology calls for falsification regardless of logical form. The theory ‘There exists a non-white swan’ isn’t falsifiable. So if you want to verify it, you first have to formulate a generalization that entails ‘There exists a non-white swan’. Then you try to falsify the new generalization. If that fails, you get among other things the desired verification. Thus the logical equivalence that Popper’s critics constantly cite is simply irrelevant. When you falsify ‘All swans are white’ you merely accept that there exists a non-white swan. You haven’t verified anything. There is logical symmetry but methodological asymmetry. Typically, you won’t even have enough information to verify ‘There exists a non-white swan’ until you’ve accepted it. So falsifying ‘All swans are white’ cannot possibly be the same as verifying ‘There exists a non-white swan’.
Falsification is not verification by another name. Don’t let the critics take the word away from you. They don’t know what they’re talking about.
I would rather talk about the critical approach to problem solving, which means you have to start off by stating the problem that you are talking about, and maybe why it matters, and then refer to the various forms of criticism that can be brought to bear on tentative solutions. In this context empirical tests are just one of the forms of criticism that can be used.
Too much discussion of philosophical and methodological issues proceeds in isolation from substantive or real-world problems, so it has the big quesation “who cares?” or “what difference does it make?” hanging over it. Working economists must be fed up with the endless discussion of MSPR and paradigms that went on for about 20 years without any resolution that I can see. Authors need to say what difference their method makes for the daily activities of scientists, including economists.
Rafe,
It is unfortunately all too easy to mix up ‘falsification’ with ‘falsifiability’. When that happens, it looks like the term ‘falsification’ is way too narrow, because it would apply only to theories that are falsifiable. Obviously, there are lots of unfalsifiable theories around that we would like to criticize; and so it seems more appropriate to talk about a critical approach than to talk about falsification. The problem with this, however, is that, if you allow falsification only for falsifiable theories, the whole of critical rationalism collapses: you’ve conceded the whole argument to the critics. Why? Because in that case falsificationism devolves into naive falsificationism, according to which a falsification is purely a matter of logic. But if that’s the case, then falsificationism really is just upside-down verificationism, i.e. Popper was an inductivist, and all the Salmons and Grunbaums are right.
On the other hand, if you recognize that, as Popper and Miller emphasize, falsification is never purely a matter of logic but must always appeal to methodological rules, it begins to look like falsification is not just for falsifiable theories. When we falsify a theory, we merely classify it as false. We have not determined that it is in fact false. We hope we get it right, but there are no guarantees. Can we not classify an unfalsifiable theory as false on the basis of methodology? Popper effectively says Yes when he issues a methodological rule against accepting stray basic statements. Through criticism and methodological rules we decide that an unfalsifiable theory is false. We falsify it.
So if ‘falsification’ seems too narrow when compared to ‘criticism’, it is only because we have mistakenly made it so. Without perhaps intending to do so, Miller makes this abundantly clear in his essay ‘What Do Arguments Achieve?’. In fact, this whole aspect of critical rationalism is latent in Popper’s work but has been sufficiently emphasized only by David Miller (Bartley started down that path but was tragically cut short). Critical rationalists are making a huge mistake when they swallow the rhetoric of propagandists like Larry Tapper about Bartley and Miller being radical loonies. It is not that Bartley and Miller are radicals. It is rather that other critical rationalists by name espouse a hybrid philosophy, an unholy alliance of a few Popperian ideas with the justificationist philosophy Popper intended to replace.
The view that ‘falsification’ is too narrow amounts to giving up the whole game. I wouldn’t do it.
Ken,
The logical equivalence between verification and falsification only concerns finite existentials or circumscribed universals. With regard to uncircumscribed universals the matter is quite different — both verifications and falsifications can only imply the falsity of uncircumscribed universals. While there are important methodological — or extralogical — issues, the asymmetry between verification and falsification manifests itself at the logical level.
Lee,
Do you have a reference for that? I would be interested in either reading or re-reading that material.
Off hand, I do not disagree with what you say. It only seems to add fuel to the fire I’m trying to light, i.e., don’t back off from ‘falsificationism’. Even if one accepts direct verification by observation, there is a methodological asymmetry that is persistently overlooked by those who harp on logical symmetry. The methodological asymmetry is this: verifications assume what they verify, and are circular; falsifications assume what they falsify, and are not circular.
The point I’m trying to emphasize is that critical rationalism replaces epistemology with methodology. The critics have not picked up on that, for if they did the logical symmetries they keep touting would be recognized as completely irrelevant. Irrelevant!
Also, in this connection, I’d like to recommend again a totally fascinating essay by Miller. It’s chapter 13 in Out of Error: Making Falsification Bite. It is a joy to read an essay that discusses a real difficulty for critical rationalism as opposed to all the hackneyed, wrong-headed rubbish repeated endlessly by critics for the last 40 years. At last! Something that actually moves forward!!
The term “falsifiable” should be reserved to demarcate propositions that can be empirically tested from those which cannot, i.e. a proposition is falsifiable when it is logically possible for an observation to entail its falsity.
The first class of propositions which are unfalsifiable are those with a tautological form or analytic character, because their falsity is contradictory. If an empirical test seems to falsify such a proposition, then something is amiss with how the test has been interpreted, or the proposition is not actually tautological or analytic.
The second class of propositions which are unfalsifiable have neither a tautological form nor analytic character, because their falsity is coherent. Such propositions are instead presupposed by the acts of observation and experiment themselves. For example, the statement “there exists an observer” is not a purely logical truth, but clearly one cannot falsify it unless one can also observe. An apparent falsification of the existence of observers would suggest that no observation exists. But if no observation exists, then nothing has been falsified. This class of propositions corresponds to something like a priori synthetic knowledge, but without the subjectivism.
The third class of propositions which are unfalsifiable are dependent on a knowledge context. For example, the conjecture that our universe is expanding, or that fusion is occuring in the core of the sun may seem empirically untestable when considered in an impoverished knowledge context. The man on the street often scoffs at scientists who claim that birds evolved from dinosaurs or talk about what chemical reactions occurred half a second after the big bang. Whether such hypotheses and propositions appear to be falsifiable depends on how one understands their place in the universe and how experiences are interpreted. Falsifiable propositions may exist that presently appear unfalsifiable in the context of our mistaken or inadequate knowledge, and one day this error may be corrected.
The fourth and final class of unfalsifiable propositions are those that uneconomical or technologically impossible to test.
Making an idea falsifiable is a tremendous achievement, because it opens up new possibilities for criticism.
Ken,
I don’t have a reference, unless LoSD counts. Anyway, it’s not particularly complicated or controversial.
To falsify ‘all swans are white’ one must verify ‘some swans are non-white’. Thus falsification “depends” on verification. But this misses the point. Only existentials need to be verified in falsificationism, whereas the traditional problem of positivism was the verification of universals.
Propositions of the form “all x are y” are unverifiable.
Propositions of the form “some x are y” are unfalsifiable.
Inductivists are continually trying to show how successive observations of “some x are y” proves, makes credible, or increases the probability of “all x are y”. But do falsificationists suggest that successive observations of “not all x are y” disproves, make less credible, or decrease the probability of “some x are y”? There is no such symmetry, because falsificationists are not presented with the same problem as inductivists, i.e. verifying universals.
Lee,
Falsificationists don’t have to try to disprove, make less credible, or decrease the probability of “some x are y” in order to provoke the charge of overlooking an important logical symmetry. All they have to do is falsify “all x are y”, for this is allegedly the same as verifying the negation of “all x are y”, i.e., the same as saying that there exists some x which is not-y. It is true as you indicate that this does not make falsificationism comparable to inductivism. It is also true that it doesn’t obliterate the asymmetry between universal and existential statements made famous by Popper in LSD. But the critics don’t deny any of that. Their point is rather that talk about falsification is nothing more than talk, because you might as easily talk about verifying the negations of universal generalizations. In other words, there’s nothing really different or special about falsificationism. It’s just verificationism turned on its head.
Now, you haven’t escaped from this criticism and I don’t think you can, because it is correct and it’s just a matter of logic. In other words, if all you’re looking at is the logic of statements, you’re definitely going to lose the argument with the critics. But that’s precisely why they want to focus exclusively on the logic of statements! In fact, the more egregious members of that club refuse point blank to discuss methodology .. what they say, in effect, is something like “oh pooh pooh … if all you’re going to do is talk about method, don’t bother me; I don’t care about it. After all, we’re all big boys and girls — I think we can figure out what to do.” End of discussion (I’ve had this discussion more than once; I know .. this is no exaggeration.)
The only successful way of addressing this criticism that I know of is to recognize that critical rationalism is about method. Falsificationism is methodologically unlike verificationism, because verifications are circular whereas falsifications are not. I really do not see how anyone can defend critical rationalism on the grounds of logic alone. And, while I do not want to put words into anyone’s mouth, I think that assessment of the situation is consistent with what both Popper and Miller have written. I’m not so sure it is consistent with what many others have written.
Once the confusion between logic and method is ejected, the label “falsificationism” because unproblematic. So .. if you want to eject it, go ahead. But don’t let the critics run you off that property against your will. You may end up paying for it later anyway.
Ken,
Falsifying “all swans are white” does verify its negation. But the negation of “all swans are white” is an existential — the logical asymmetry that Popper spoke of concerns the status of universals. That existentials can be verified is uncontroversial* and irrelevant. Certainly one might describe falsifying as verifying the negation of a universal, but since the negation of a universal is an existential, one could equivalently, and with less verbosity, describe falsifying as verifying an existential. But this is exactly what was said in the first place.
This is the logic of the situation:
Existential – verifiable but unfalsifiable
Universal – falsifiable but unverifiable
With regard to universals, there is an aymmetry between verifiability and falsifiability. The logic of falsification only requires verification of existentials. Among the logical content of existentials are negated universals, because negated universals are just another way of stating an existential.
*Of course, any verification is theory-laden and fallible.
In a way, falsificationists are more consistent verificationists than verificationists. Each verifies existentials (i.e. negated universals) and nothing more. But while falsificationists are satisfied with the logical relationship between existentials and universals, verificationists are not. They continually search for a means to include universals (i.e. negated existentials) among the logical consequences of existentials.
Lee,
Regarding the logic of the situation — Yes, that’s the asymmetry Popper features, which as I noted too is perfectly fine. Agree with that. But the first sentence of your last post is precisely what I want to deny. Falsificationists are not verificationists, and the reason can only be methodological because, as we both agree, every time you falsify you also verify. Now you suggest that verificationists “continually search for a means to include universals (i.e., negated existentials) among the logical consequences of existentials. That strikes me as a good idea. But I would like to put some flesh on it along the lines suggested by Rafe. In other words, I think it is incumbent upon critical rationalists to say exactly what scientists should do or do differently. What practical difference does it make to their work?
One idea I’ve had along these lines is that the tragic case of Rusi Teleyarkhan may provide some insight. There was at one point a video on-line that went through the history of his case. I cannot find it at the moment, but I’m still looking. As some may know, Teleyarkhan has been reprimanded for misconduct. My theory is that verificationism is what got him in trouble. This is not to pass judgment on the truth of his theory. For all I know, he is right, though a majority of scientists seem to think he’s wrong (and it would be really interesting if Teleyarkhan turn out to be right). What I suggest is that people view his story and make suggestions about exactly what he, perhaps, should not have done or said — aside from the obvious charge of falsifying data or putting a name on a research paper that didn’t belong there. I think it quite possible that Teleyarkhan’s behavior may bear out your contention. I’m not sure. If so, I think it could be put in some concrete terms.
If anyone can find the on-line video about Rusi Teleyarkhan and cold fusion, please post it.
Here it is:
http://video.google.com/videoplay?docid=8522152695084736832#
Since nobody seems to be responding here I’m going to propose a methodological principle which I think may be drawn as a lesson from the case of Rusi Teleyarkhan. This anti-verificationist principle is very close to what Lee said above:
a) Observational and experimental evidence may not be interpreted as support for any claim of greater logical strength than an existential hypothesis.
It should be clear from the video that Teleyarkhan violated this principle, even though his theory may be correct. My guess is that this is where the trouble for Teleyarkhan started, and that his career might still be intact if he had not done this. I don’t think that the mistake was dishonest; I think it was just a mistake, though I also think it plausible that Teleyarkhan may have been less than honest later on when lots of pressure had been brought to bear.
The anti-verificationist principle above pairs rather nicely with the principle of empiricism, which may also be accepted:
b) No hypothesis that has been accepted may be rejected except on the basis of observation or experiment.
I don’t think that either one of these principles has any obvious implications for Austrian economists. But there may be other methodological principles that do. Any ideas? Again, this is in line with the question implied by Rafe. What actual, practical proposals would critical rationalists make when it comes to methodology?
BTW, the need for another bylaw along more “social” lines may be noticed in the Teleyarkhan story. What is it?
Methodology should be objective. I think Matt noted that recently. Popper tended to follow this principle.
Both praxeologists and Bayesians tend to create prescriptive rules about what one can or should feel about hypotheses given particular circumstances — this is more like ethics than methodology. But even Popper’s criticism of dogmatism was conducted in an objective manner. The methodological rule against reinforced dogmatisms concerned the objective logical structure of hypotheses and not subjective dispositions. Popper once suggested that a little dogmatism (in the sense of subjective conviction or certainty) may be a good for the social enterprise of science, because all hypotheses deserve thorough and motivated defence. But this should not be confused with advocating hypotheses whose logical structure is such that no criticism can break its layers of deflective stratagies. One may adopt unintentionally adopt a reinforced dogmatism while having a thoroughly critical attitude, because its dogmatic logical structure may go unnoticed.
Lee,
Yes, I agree about the objectivity. So, for instance, regardless of how RT felt about the significance of the data, he should not have presented it in a paper as evidence for cold fusion, because all he had were existential statements and the cold fusion hypothesis is logically stronger than that. Do you think the principle adequately meets the criterion of objectivity?
Listen carefully to what Martin Fleischmann says and you will notice an obvious fault in the scientific community. That needs to be addressed.
So, in other words, RT could, if he wanted, publish a paper reporting his results, and summarize those results by noting that he failed to find any evidence to rule out cold fusion. That would have been a different claim entirely, and he would not have violated the principle. All would be fine. Others could come in an rule out cold fusion with no problem.
Then also the editor at Science should have blocked the paper for the same reason. He should have sent it back saying “Withdraw the claim of support for cold fusion. Just report that you failed to find any evidence to rule that out.” Again, all would have been fine. It’s verificationism that caused this thing to blow up.
Gee … what about the fact that these scientists have all had their careers ruined because they proposed theories that were subsequently rejected. Doesn’t anyone see even the slightest problem with that from the point of view of critical rationalism?
???
Here is a further refinement, and a connection back to Austrian economics.
The principle proposed above can be brought into line with the Popper-Miller proof of 1983 reformulating as follows:
1) Observational or experimental evidence may not be interpreted as support for any hypothesis that is deductively independent of the evidence itself.
The cold fusion hypothesis is deductively independent of the evidence presented by Teleyarkhan. Thus it is a conjecture. Teleyarkhan said: “the evidence speaks for itself”. But this is wrong. The evidence never speaks for itself. It must always be interpreted. What the principle here says is that, while evidence must always be interpreted, there is one interpretation in particular that is not allowed. The problem with Teleyarkhan’s work is not that he proposed the possibility of cold fusion, but that he interpreted the data as support for the theory. If he had not done that, he would have simply stated that he found no evidence to rule out cold fusion, which should not have caused any kind of scandal.
The Austrian economists, on the other hand, claim that empirical evidence is irrelevant to the formulation of economic theory. They do, however, endorse Ricardo’s Law of Comparative Advantage. Mises called it the Law of Association, about which Gene Callahan says: “Although the initial application of the law was to trade, it is a universal law applying to all human cooperation.”
The challenge for Austrian economists then is clear: to demonstrate how any such universal law is logically entailed by any of their so-called actions. There is no book or publication I know of by any Austrian economist that accomplishes this. And in fact I think anyone who has a rudimentary knowledge of modern logic should realize straightaway that it cannot be done. Still, Austrian economists carry on, repeating this assertion over and over again. Their publications invariably do nothing more than set forth a number of supposed axioms, and then claim that the economic theories follow deductively without giving the slightest hint about how the deduction can actually be carried out. Nor do Austrian economists seem to understand that, if their axioms were logical tautologies as they sometimes claim, and if they did actually entail an economic theory, that theory would necessarily be devoid of empirical content.
In any event, it seems to me high time for Austrian economists to either put up or shut up on the the topic of methodology: show us the deduction explicitly; name the logical rules of inference invoked at each step. My prediction is that 1) nobody will be able to do this, and 2) no Austrian economist who currently claims that it can be done will under any circumstances change their opinion in this respect. It is simply a dogma, an article of faith. The amusing thing about this particular article of faith is that it isn’t even consistent with rudimentary modern logic.
The plea has been made in this thread to get away from abstract discussion and get down to matters of scientific practice. What I would like to know is what practical difference it could possibly make to Austrian economists if all of their so-called axioms were simply discarded. If the law of association was regarded as a conjecture, and not as the conclusion of a deductively valid inference (which it is not), what difference would it make?