PROBLEMS WITH POPPER SCHOLARSHIP

See also Trashing Popper.

Daniel Hausman The Inexact and Separate Science of Economics (1992)

In this book Hausman defends the old-fashioned view that economics proceeds by deducing the consequences of  “apriori” axioms in particular situations. (This could be called Situational Analysis). The axioms are “inexact”, and they only permit the prediction of tendencies (propensities?). He suggests that confidence in the implications (predictions and explanations) of economics comes from confidence in the axioms rather than from testing their implications.

An initial reaction, how realistic is it to suggest that our confidence in the starting point of deductions has nothing to do with the results (the implications), even if the process is not depicted in terms of testing. Every economic action can be regarded as an experiment, even if it is just the process of living without any formal chain of reasoning and a detailed analysis of the results.

I suppose the axioms that he has in mind would be such things as the laws of supply and demand, diminishing returns, marginal costs, opportunity costs, comparative advantage, and the benefits of the division of labour.

Some texts such as Human Action by von Mises give the impression that everyone has to learn economics from scratch, perhaps by meditating on the concept of human action, but of course we inherit a body of economic principles, methods and precepts. The purpose of scholarship, research and critical thinking is to improve them.

Hausman’s valid point is that the old-time economists did not do too badly and the efforts that have been made in the 20th century to improve economics by injecting modern “state of the art” epistemology and methodology have done nothing  to help. He is especially critical of the positivists, the Popperians and the post-positivists.

To signal the bottom line of my argument, criticisms of Popper for the most part do not hit the target, especially the criticisms that are directed against Popperian "falsificationism". As I will demonstrate elsewhwere, that perception of Popper is false and misleading. Following the lead of Kuhn and Lakatos that is the standard treatment of Popper. The argument proceeds as follows. First, falsificationism is defined as the doctrine that theories (or hypotheses) can be decisively falsified. Then it is pointed out that (1) this doctrine is logically incoherent (due to the Duhem effect, for example), (2) it is not practiced by scientists and (3) it would be a bad thing if they tried to do so because they would have to discard theories and hypotheses when they are apparently refuted, even if there are no better alternatives.

A short response to those criticisms: (1) Popper did not claim that hypotheses can be decisively falsified. See the extract from The Logic of Scientific Discovery at the end of this paper. (2) Scientists who are not prepared to test their theories, or generally adopt a critical attitude, are asking for trouble, if they don't test their theories someone else will eventually do so and the sooner problems are identified the better. (3) He appreciated that an apparently refuted hypothesis need not be immediately discarded (for several good reasons). Critics have simply not paid attention to Popper's counter-arguments.

This is a very strange state of affairs, where this unhelpful conception of Popper's ideas is so widespread that apparently people have given up reading the original texts. A review that suggested that this reading could  be important received mostly "unhelpful" votes on the Amazon website. One of the most influential contributors to this misperception of Popper is probably Alan Chalmers in his otherwise excellent book What is this thing called science? My critical view and also positive comments can be found in this review.

This reading started as an exercise in documenting some problems with Popper scholarship. It will help to put both Hausman’s argument and Popper’s ideas in a better light to read or re-read Popper not as a falsificationist, obsessed with the demarcation between science and non-science, but as a fallible apriorist (in the language of Barry Smith), concerned to promote the search for the truth and the growth of knowledge by making the best use of our imaginative and critical faculties (and evidence) to solve problems of all kinds (theory, method, practice).

Chapter 9 in Hausman's book describes the methodological revolution of the 20th century when the deductive method of the 19th century came under serious attack in the 1930s. These attacks came from the logical positivists, from people who picked up bits of Popper (notably the idea of falsification), from Samuelson’s version of positivism (conventionalism?) and from Friedman’s instrumentalism.

He quotes from Koopmans.

"The theories that have become dear to us can very well stand by themselves as an impressive and highly valuable system of deductive thought, erected on a few premises that seem to be well-chosen first approximations to a complicated reality. They exhibit in a striking manner the power of deductive reasoning in drawing conclusions which…are highly relevant to questions of economic policy. In many cases the knowledge these deductions yield is the best we have, either because better approximations have not been secured at the level of premises or because comparable reasoning from premises realized as more realistic has not been completed or has not been found possible. Is any stronger defense needed, or even desirable?”

Hausman is sympathetic to that view but he suggests that more is needed to flesh out the scientific acceptability of “well chosen first approximations”, to identify the role of observations and experiments, and to legitimate that approach in the face of “another three decades of misconceived methodological debate largely divorced from methodological practice”.  The main villains in this debate were Popper and Lakatos and their interpreters because the positivists were mostly sidelined as a result of the post-positivist revolt of the 1960s and 1970s. (Though the corrosive legacy of the positivist-empiricists lives on in the US schools of philosophy and in the customary misinterpretation of Popper and the alleged failuresa of “falsificationism”).

Hausman concluded that “Popper’s and Lakatos’ views are no more hospitable to towards the deductive method than are those of the logical positivists and logical empiricists.” And if Popper is correct then the traditional approach affirmed by Koopmans is “completely untenable”.

It is interesting to note that inductivists sometimes mock Popper’s “deductivism”!

Moving on to the critique of Popper in Chapter 10, Hausman notes that his criticisms are widely accepted among philosophers (citing Levison  (1974), Lieberson (1982a, 1982b), Putnam in the 1974 Schilpp volume, Grunbaum (1976)and Salmon (1981). There are a number of standard critiques of Popper in the literature, there are also standard counter-arguments which to my mind are devastating, though they appear to be ignored by the critics. This is a good example, showing how various arguments taken from the history of science fail to undermine the logic of Popper's situation. The paper by Putnam (1974) is especially disappointing. He claimed that Popper did not take account of additional theories or statements that are required to test a general theory. These are the statements that Popper described as the initial conditions for the test observations. The criticism is a non-sequitur. Salmon's argument is briefly cricicised here.

10.1 Demarcation

It is unfortunate that Popper placed so much emphasis on the demarcation problem. He regarded this as one of the two fundamental problems in the philosophy of science and I am not sure why this is the case. It may be a legacy of Newton’s triumph which inspired a quasi-mystical attitude towards science which replaced the older concept of science as any organized body of knowledge, while to be scientific was simply to be deliberate and systematic in addressing any subject, from angling to astronomy.

Unfortunately Popper started his career in dialogue with the positivists who were obsessed with the methods of Science (and the justification of belief) and also the demarcation between science and metaphysics. Consequently he tried to show them a way out of the rut of verification and justification (though without success). As a result of his prolonged engagement with the positivists it was widely perceived that the central core of his thinking was “falsificationism” because that was the most obvious difference between his position and that of the positivists. In fact he took a number of epistemological turns - the "conjectural turn", the "objective turn" (from subjective beliefs), the "social turn" (to the rules of the game and institutional analysis) and the metaphysical turn.  It is important to appreciate these "turns" to get a grip on his thinking.

Moving on to Hausman and the problem of demarcation, two issues appear. First, he was confused by the distinction that Popper drew between the logic of falsifiability and the practice of testing (attempted falsification). And second, like all the other philosophers, especially those who cleave to the logic of induction in criticism of Popper, Hausman insists on the crucial importance of  positively justified beliefs, so scientific theories need to have a stronger warrant than merely standing up to criticism.

On falsifiability he wrote: “I cannot judge how important logical falsifiability is to Popper, since he apparently contradicts himself on this matter.”

That perception of contradiction is simply based on failure to appreciate the distinction that Popper drew between (logical) falsifiability and (actual) falsification.

To be fair, it is easy to pick up the idea that Popper is primarily concerned with the scientific status of theories, as opposed to practices, and maybe people who expound his ideas need to be more clear about this. Certainly some of my early essays did not do so.

Hausman seems to conflate the distinction between falsifiable and unfalsifiable statements with the difference between science and non-science, which of course raises the issue – what do we mean by science?  That has to be read in context to identify whether the issue relates to particular statements, the corpus of science at large, the methods of science or the great human enterprise of science with all its social and political aspects.

When Popper was wearing his "logic of science" hat, he used the logical asymmetry of verification and falsification to make the point out that the positivists and inductivists do not understand the uses and especially the limitations of evidence. They falsely think that there is some way to provide inductive support for a hypothesis, perhaps even a numerical probability value. He insists that the logic of scientific investigation is the dedictive logic of the Modus Tollens. And so the answer to the question, “what can you do with a true existential statement?” is this: you can refute a universal statement but you can’t verify a universal statement.

Moving from the logic of testing to the practice, Popper emphasized that there can be no decisive falsification due to the Duhem problem, to the uncertainty of evidence, and the possibility that test statements can simply be ignored or dismissed or explained away by ad hoc hypotheses. At this point Popper put on his "rules of the game" hat to formulate conventions, or rules of the game of science that will maximize the effectiveness of criticism, especially the criticism of testing by evidence.

A recent book by Jarvie clearly explains  a point that I was groping towards for some time, that the unifying feature of Popper’s work on science and  society is the all-important role of conventions, the rules of the game.

Popper’s approach focuses attention on the comparative merits of rival theories, or perhaps the problem of preference (instead of assessment) which Hausman (in another paper) identified as the centre of attention in the field after Popper and other distractions were put aside. This more "advanced" position could have been derived from a  more sympathetic reading of Popper from the very beginning and hence averted the large volume of unhelpful "Popper bashing" over the year. 

10.2 Logical falsifiability and Popper’s solution to the problem of induction

In this section Hausman argues against Popper’s solution to the problem of induction. He notes that Popper’s proposal for the demarcation criterion gave him the clue to solving the problem of induction, at least the problem as it impacts working scientists. They are mostly trying to get on with the job of inventing better theories, unlike the inductivists (aparently including Hausman) who are looking for the justification of beliefs. The problem for working scientists is how to use evidence most effectively to promote the growth of knowledge (and to engage in any kind of problem-solving). Popper’s answer is to use evidence for testing, as one of the various forms of criticism.

By testing and by finding refutations, we are supposed to advance by eliminating error. Actually I am not happy with that formulation, we really progress by inventing better theories; theories which explain more, unify disciplines, predict more precisely (if it matters) and stand up to criticism, including the test of practical application in technology or experiments.

Hausman writes:

"Popper is proposing to cut the linkage between knowledge and justification altogether. Conjectures about the world constitute knowledge if they are true. In attempting to falsify them, scientists sometimes find out that they are false  and not knowledge at all. That which has not been falsified one takes to be knowledge. Justification has no role. Hume is correct that justification is not available, but Popper maintains that it is not needed either. I cannot accept this dramatic problem shift."

That passage contains a very helpful insight into the mind of the justificationist. There is no concept of conjectural knowledge. It it is false it is not knowledge. So when Newton’s theory was refuted it ceased to be knowledge? So much for the achievement of Newton!

Hausman insists that there must be positive reasons to support a theory, not just the negative reason that it has not yet been refuted.

"One has no better reason to expect that the predictions of well-tested theories will be correct than to expect that untested theories will predict correctly. This skeptical view is central to Popper’s philosophy of science. For, if Popper were to admit that the results of testing can give one reason to believe in the reliability or approximate truth of claims about the world, then he would have to face the problem of induction and attempt to explain how they can do so. Popper’s option is simpler but it fails."

First, we are still waiting for the inductivists to explain how evidence supports theries in the strong, positive sense that they demand. Secondly, Popper's "conjectural turn", his "non-justificationism" shifts the focus from the positive justification of true beliefs to the formation (and revision) of critical preferences between alternative theories. Popper has written that we can often justify a preference in a way that we cannot positively justify any theory. Forming the preference is a matter of situational analysis, rather like making choices of goods or investments in the marketplace.

To paraphrase the celebrated section of Orwell’s Animal Farm, “all theories are conjectural but some are more conjectural than others”. The original reads “All animals are equal but some are more equal than others”.

I should say that some theories are better than others, that is a matter of the situation, their track record and the nature of the problem – theories that are known to be false may be good enough for calculations (as Koopmans noted above).

Appealing to the track record is not smuggling in a principle of induction, it merely appeals to the existence of regularities in the world, that is a metaphysical theory about the world and nothing to do with the logic of induction. This "regularies" principle is often used as the last resort of the inductivist after appeals to inductive logic have failed, then the regularity of the universe is postulated as the inductive principle.

Hausman continued with several pages of criticism of falsificationism as a methodology,. This appears to be damning for Popper but in fact it beside the point because Popper’s views on falsification apply to the use and the limitations of of evidence and the concept of falsification cannot be viewed as the whole of his methodology which is essentially about making use of all the forms of criticism to advance the critical discussion of theories.

This chapter is followed by a critique of  Lakatos who he depicts as a rather more interesting kind of  Popperian but also useless for economists.

To demonstrate the durability of lazy Popper scholarship, see Hausman’s Introduction to The Philosophy of Economics (CUP).

This collection of papers appeared in 1984, with revised editions in 1994 and 2008. The Introduction runs through some of the familiar territory about assessment and demarcation, Hume’s problem, Neurath’s metaphor of  the ship that has to be rebuilt while at sea, a reference to Carnap on inductive logic, “a failure that helped to locate more promising developments in confirmation theory”!!

He then came to Popper.

After sketching Karl Popper’s “more radical” views on induction he replicated his confused comments on Popper’s demarcation principle. He then went on to Popper’s rules of the scientific game. “Popper is often concerned instead to distinguish those attitudes, rules and practices that distinguish a scientific community from other attitudes and practices.”

But then in the 1984 edition he wrote :

"According to Popper, scientists propose bold conjectures and then seek out the hardest possible tests of them. When the conjectures fail those tests, no excuses are permitted. The theories are regarded as refuted, and new conjectures are proposed and scrutinized. (31)"

Note 31 is Conjectures and Refutations pp 49-52, though I can’t see anything there about immediately discarding refuted conjectures and turning to new ones.

It is not surprising that there is no supporting citation from Popper because he was aware that apparent refutations can be contested and he was never a naïve falsificationist, contra Lakatos and Kuhn. He appreciated that new theories need to be developed to get over early problems, one of his rules of procedure was that no theory should be dropped without good reason, such as the availability of a better theory and he even suggested a methodological excuse for a whiff of dogmatism to allow time to develop new theories. In his contribution to the collection Criticism and the Growth of Knowledge (1974) Popper wrote “I have always stressed the need for some dogmatism: the dogmatic scientist has an important role to play. If we give in to criticism too easily we will never find where the real power of our theories lies”. Bartley disputed this formulation because it is enough to signal that adverse results render a theory “problematic” and no hint of dogmatism is required to keep the theory under consideration for development on a conjectural basis (like every other theory).

The same “no excuses” passage stands unchanged in the second edition of the book.

In the third edition there is a minor change.

"These rules require that when the conjectures fail those tests, scientists do not make excuses. Instead they should regard the theories as refuted and they should then propose and scrutinise new conjectures. (20)  As many have noted, including Thomas Kuhn and Imre Lakatos, it is a good thing that scientists do not follow these rules."

Note 20 refers to The Logic of Scientific Discovery chapter 5. The statement in the text does not convey Popper’s nuanced position regarding putative refutations and chapter 5 "The Problem of the Empirical Base" does not address the issue of responding to refutations at all.

This is very strange. A clearly false interpretation of Popper’s ideas has remained in place for 25 years through two revised editions of the book and presumably during this time the Introduction has been read by a large number of scholars with an interest in the philosophy of science. But apparently it has not been read by a single one who knew enough about Popper to recognise the mistake.  In the latest edition note 42 of the Introduction acknowledges the assistance of three other scholars who read drafts of the Introduction and improved it. These are Wade Hands, Kevin Hoover and Margaret Schabas. “One of great (sic) privileges of having worked so long on economic methodology is being able to count such wonderful people and wonderful intellects as friends”.

Wade Hands in recent times has started to come to grips with the real Popper rather than the Popper of the Received View which depicts him as an eccentric positivist. Hence the expectation that the book following Reflection Without Rules would be shorter and more helpful.

The note also nominates a platoon of scholars who Hausman recruited to help with the book. Did they also read the Introduction and miss the false interpretation of Popper? Or is it generally accepted that misrepresentation of Popper does not matter? Or that it is unhelpful to draw attention to these things? That is the verdict of several readers of an Amazon review of a recent collection of papers in the history and philosophy of science, where it is suggested that students should be invited to read the original  works to check whether the editorial critique of Popper is valid.

Footnote

An extract from The Logic of Scientific Discovery to demonstrate that Popper cannot be regarded as a naive falsificationist.

"I am quite ready to admit that there is a need for a purely logical analysis of theories, for an analysis which takes no account of how they change and develop. But this kind of analysis does not elucidate those aspects of the empirical sciences which I, for one, so highly prize. A system such as classical mechanics may be 'scientific' to any degree you like; but those who uphold it dogmatically--believing, perhaps, that it is their business to defend such a successful system against criticism as long as it is not conclusively disproved--are adopting the very reverse of that critical attitude which in my view is the proper one for the scientist. In point of fact, no conclusive disproof of a theory can ever be produced; for it is always possible to say that the experimental results are not reliable, or that the discrepancies which are asserted to exist between the experimental results and the theory are only apparent and that they will disappear with the advance of our understanding. (In the struggle against Einstein, both these arguments were often used in support of Newtonian mechanics, and similar arguments abound in the field of the social sciences.) If you insist on strict proof (or strict disproof *1 in the empirical sciences, you will never benefit from experience, and never learn from it how wrong you are."  [My emphasis, RC]

"*1  I have now here added in brackets the words 'or strict disproof' to the text (a) because they are clearly implied by what is said immediately before ('no conclusive disproof of a theory can ever be produced'), and (b) because I have been constantly misinterpreted as upholding a criterion (and moreover one of meaning rather than of demarcation) based upon a doctrine of 'complete' or 'conclusive' falsifiability.) "

From Section 9. Why Methodological Decisions are Indispensable.  Page 51.


Entry to the siteFull index of papers on site       More "trashing Popper" papers