Tuesday, 21 March 2017

Some brief notes surrounding popular misconceptions of the objection from ill-fit

Here are some brief notes on the perpetuation of three philosophic myths surrounding Popper's demarcation criterion:

(i) Popper's demarcation criterion sets out the limits of the natural sciences from non-science or pseudo-science.

(ii) Popper's criterion explicitly demarcates isolated scientific theories from non-scientific or pseudo-scientific theories.


(iii) Popper's criterion is either too restrictive and/or permissive a criterion to demarcate the domains of science from non-science or pseudo-science (the objection from ill-fit).

All are false.

(N.B. In fact, Ayer's 1936/46 criterion and Carnap's 1956 criterion are either trivial or reduce to Popper's 1934/35 criterion. See: Popper (1959, 65-66), Ayer (1936, 97-99) and Carnap (1956). Thus, there exist three different versions of the objection from ill-fit: the first objection addressed below is widely off the mark; a second objection I allude to has been answered in the literature (cf. Gillies, 1971); the third, seen in, e.g. Church's review of Ayer (1949), was so debilitating it put to rest most work done on the demarcation problem.)

1. Introduction

A number of philosophers dismiss a specific territorial problem of demarcation as unsolvable (Kitcher, 1982; Bunge, 1983; Thagard, 1978; Laudan, 1983). Furthermore, the shift in recent years has been towards solving a normative problem of demarcation at the expense of this territorial problem (cf. Boudry & Pigliucci, 2013). One objection is routinely advertised as a potential culprit for abandoning territorial problems. We can call it the objection from ill-fit.

This objection may be valid when directed at some possible territorial criteria; however, I argue it is spurious when presented against its primary intended target: the criterion and problem set out by one main advocate for demarcation, Sir Karl Popper. It is spurious for three reasons: (i) it misidentifies the scope of the boundaries that are drawn, (ii) what is to be demarcated, and (iii) what purpose the demarcation criterion serves. In contrast to popular myth, Popper’s criterion sets out the limits of empirical inquiry, not the natural sciences; it applies to theoretical systems, not individual theories; it is taxonomic, not normative.

I then explain how the objection from ill-fit looks prima facie promising, but is likely based on a mistake in exegesis. This has lead to the entrenchment of the philosophical myth that the objection from ill-fit in this form targets Popper’s demarcation criterion.

(N.B. A similar objection from ill-fit, however, may be more fruitful: it targets Popper's 1934/35 criterion, as well as A.J. Ayer's 1936 and Rudolf Carnap's 1956 respective criteria, since all three criteria exclude universalised probabilistic statements. Carnap and Popper both present replies to this objection that fold in their normative criteria. The problem is that, as Carnap notes, universalised probabilistic statements 'are neither falsifiable nor verifiable… neither their negations nor they themselves are completely confirmable...' (Carnap, 1937, 27; cf. Popper, 1959, 190). However, both Carnap and Popper provide methodological solutions to this problem (e.g. Popper, 1959, ch. VIII). This more robust objection from ill-fit is not the end for demarcation criteria: a more rigorous defence is available in Donald Gillies' 1971 article, A Falsifying Rule for Probability Statements.)

Moving focus back to these philosophic myths, I set out a reconstruction of a number of unfortunate events that plausibly lead to the present state of affairs: The Logic of Scientific Discovery (1959) is the English translation of Popper’s 1934 book, Logik der Forschung. Due to a quirk of the German language that does not exist in English (i.e. the respective scopes of ‘science’ vs ‘Wissenschaft’) Popper chose to translate a key technical term in the German edition, ‘empirischen wissenschaft’ as ‘empirical science’, rather than ‘knowledge gained by experience’ or any other suitable translation. Popper then failed to adequately signpost the meaning of this technical term.

Subsequently, many of his critics mistakenly took the aim of his stated demarcation criterion to delineate the boundaries between the natural sciences and other forms of empirical inquiry; other philosophers concluded the criterion was normative rather than territorial; other inattentive critics took the aim of the criterion to be limited to individual theories. These three mistakes in exegesis lead to the popularity of this version of the objection from ill-fit. I then locate the key passage that likely lead to this mistake in The Logic of Scientific Discovery. I show how Popper aimed to set out a criterion for demarcating empirically significant from empirically non-significant domains of discourse. Consequently, the objection from ill-fit is spurious when directed at Popper’s territorial criterion. I conclude that no philosopher should continue to present the objection from ill-fit as targeting this particular territorial criterion of demarcation.

2. A misunderstanding of the scope of demarcation criteria

Though brief, what follows is a fairly accurate depiction of what the philosophical community takes to be the two possible purposes of Popper’s proposed demarcation criterion: ‘Popper… wanted to distinguish scientific theories or hypotheses from nonscientific and pseudoscientific ones’ (Pigliucci, 2013, 10, cf. Preston, 1994, 320). We can call the two problems a normative and a territorial problem of demarcation.

Territorial demarcation problems are taxonomic: distinctions are made without judgment over their respective values. One formulation of a territorial problem requires elucidating ‘the distinction between science and nonscience in general’ (Mahner, 2013, 31; cf. Boudry, 2013, 81 and Nickles, 2013, 101). This formulation is repeatedly attributed to Popper in the philosophical literature. For one example, the Stanford Encyclopedia of Philosophy article on Popper states that falsifiability is a ‘criterion for demarcating science from non-science’ (Thornton, 2016).

On the other hand, the normative problem attempts to distinguish ‘bona fide science from pseudoscience’ (Boudry, 2013, 79). ‘Science’ in this sense is intended as an honorific bestowed on certain theories, practices or communities and judgment of their respective values. This formulation of the normative problem is attributed to Popper as well (cf. Laudan, 1983, 118; Boudry, 2013, 80). As I show, this depiction of Popper’s aims in setting out his version of the demarcation problem is inaccurate.

Contra Boudry (2013, 82), Mahner (2013, 114) et al., Popper’s interests are not set towards demarcating the natural sciences from non-science. In fact, he is explicit about this in the 1959 introduction to the English translation: the natural sciences are ‘common-sense knowledge writ large… Its very problems are enlargements of the problems of common-sense knowledge’ (1959, xxvi, cf. xix).

Popper’s purpose of his proposed falsifiability criterion is to ‘provide a suitable distinguishing mark of the empirical, non-metaphysical, character of a theoretical system’ (ibid., 11, cf. 14); the problem of demarcation is to find ‘a criterion... [that] would enable us to distinguish between the empirical sciences on the one hand, and ... “metaphysical” systems on the other’ (ibid., 11). Any proposed solution to this territorial problem elucidates a theory of the absolute limits of empirical inquiry, not the limits of the natural sciences.

Popper explicitly defines ‘empirical science’ (‘empirischen Wissenschaft’) as follows: ‘The system called “empirical science”... must satisfy the criterion of demarcation, … i.e. it must not be metaphysical, but must represent a world of possible experience’ (ibid., 16-17; cf. 1934, 11-12). Popper’s use in Logik of the phrase ‘empirischen Wissenschaft’ tracks the German use of this technical term. For one example, I quote from Weibl and Herdina’s English-German technical philosophical dictionary: ‘empirical science’ is defined as ‘empirische Wissenschaft … empirical knowledge, knowledge by experience, empirical generalisation’ (1997, 120). In contrast, Popper uses the technical term, ‘natural science’ (‘naturwissenschaft’), to apply to the sciences (cf. 1959, 18, 48, 58). Thus ‘empirical science’ extends to cover, for example, knowledge gained from experience about assembling an IKEA table, a demonstrably non-scientific activity.

3. A misunderstanding of the subject to be demarcated

Furthermore, Popper’s falsifiability criterion is routinely falsely purported to be the following: a theory is scientific if and only if it is falsifiable (Thornton, 2016; Nickles, 2013, 101). That is, a universal statement counts as belonging to the natural sciences if and only if it is prohibitive: it rules out the possibility of some state of affairs. If the statement is not prohibitive, it is either pseudoscientific, nonscientific or both.

Any examination of Popper’s criterion reveals it as follows: 'it must be possible for an empirical scientific system to be refuted by experience' (ibid., 18), that is, the theoretical system as a whole must rule out the possibility of some systems of statements 'of a lesser level of universality' (ibid., 25).

Furthermore, further textual evidence demonstrates Popper's repeated insistence that his falsifiability criterion is directed towards demarcating systems of statements, rather than demarcating singular theories. Popper repeatedly claims '... we can indeed falsify only systems of theories' (1983, 187; cf. 1934/5, 12-13; 1959, 18; 1963, 56, 66). Note, 'it is important to remember that [the criterion of demarcation] applies to theoretical systems rather than to statements picked out from the context of a theoretical system' (Popper, 1983, 178; cf. 1963, 186, 256).

4. The erroneous objection from ill-fit

With these two common misconceptions laid out, I will now address the popular but erroneous version of the objection from ill-fit.

Carl Hempel (1950, 1951) notes territorial criteria are bound to be both too restrictive and too permissive. While Hempel's focus is directed at Ayer and Carnap's criteria, this general method has been appropriated to apply to Popper's criterion of falsifiability.

One example arguing that falsifiability is a criterion that is too permissive comes from Larry Laudan: it 'has the untoward consequence of countenancing as "scientific" every crank claim which makes ascertainably false assertions' (1983, 121). Other examples can be found in Lakatos (1981, 117), Mahner (2013, 30) and Boudry (2013, 87).

Philip Kitcher, on the other hand, asserts it is too restrictive: 'one can appeal to naive falsificationism to show that any science is not a science' (1982, 44): a scientific theory requires auxiliary hypotheses in order to make any predictions, thus no theory qualifies as 'scientific' on its own. See also Thornton (2016) and Putnam (1974) for other examples in the literature.

If no proposed criteria can set out the boundaries between the natural sciences and non-science, philosophers then abandon this territorial problem and redouble efforts to solve normative problems of demarcation (Boudry & Pigliucci, 2013).

5. Why this formation of the objection from ill-fit is mistaken

First, it confuses normative and territorial criteria: since the territorial criterion set out by Popper delineates the absolute limits to empirical inquiry; it is not normative (1959, 14, 16). It does not matter if there exists ascertainably false or other epistemically objectionable empirically significant systems of statements that are falsifiable. This is to be expected for any solution to a territorial criterion proposing the limits to the domain of some form of discourse. In this case, empirical systems of statements that are known to be false are empirical because that they are known to be false through empirical inquiry.

An analogy illustrates the absurdity of this reasoning: it would be as fruitful to object to falsifiability as a territorial criteria on these stated ground as it would to object to criteria for synthetic statements on the grounds that under some stated criterion of synthetic statements, a synthetic statement can be ascertainably false and still be synthetic; that is, the objection from ill-fit as framed previously is not a cogent objection.

Second, the objection from Kitcher et al. that scientific theories on their own lack any predictive capacity is also predicated on a mistake in exegesis: the purpose of the stated demarcation criterion is not to set out what makes a theory empirically significant; it is a criterion for the empirical significance for theoretical systems not reducible to systems of statements 'of a lesser level of universality' (ibid., 25).

6. A plausible explanation for the origin of the objection from ill-fit

The explanation for the prevalence of the objection from ill-fit as presently framed is, I think, due to a problem in translation. The translators of the English edition–Popper, Julius and Lan Freed–set out to be faithful to the text. From the introduction to the 1959 translation, we have the following: 'The original text of 1934 has been left unchanged for the purpose of the translation' (ibid., xiv).

This aim in translation is impossible: a change to a text is necessary in translation. Examining the original edition and translation, there is a strict adherence to sentence-by-sentence translation in the first chapter. What has not changed, however, is the order of sentences. Furthermore, whenever possible a 'literal' one-to-one English equivalent of a German term is given.

This choice in translation sets the reader up for confusion: in the first sentence of the opening chapter, Popper introduces a technical term ('empirical science') that implies through the use of the term ‘science’ some important relationship with the natural sciences. Rather than define his technical term on the first page in an added footnote, the definiens is introduced after he had used the term eighteen times (ibid., 3-4, 9, 11-12, 14-16).

Furthermore, this problem is compounded: in these sixteen pages before the definiens, Popper frequently uses the terms 'scientific' and 'natural sciences' in ways that repeatedly invite equivocation if the reader is not aware that they are each themselves technical terms in German philosophical discourse, and reflected as such in the German edition. Popper’s use tracks the use of their corresponding German technical terms, in which the closest analogue in English to the word Wissenschaft is the catchall term 'science'. The only differentiation made to each term is the choice to translate these key terms by amending ‘empirical’ ('empirischen') or 'natural' ('natur') to 'science' ('Wissenschaft'); their technical meaning is not signposted by Popper.

This produces exegetical confusion in the very passage first setting out the demarcation criterion: the German edition says, 'Ein empirisch-wissenschaftliches System muß an der Erahrung scheitern können' (1934, 13). Popper translates 'empirisch-wissenschaftliches System' as ‘an empirical scientific system’ (1959, 18). This can be understood as either ‘an empirical-scientific system’, i.e. as a system of empirical knowledge, or ‘an empirical scientific-system’, i.e. as a system of theories in the natural sciences.

Due to this choice in translation and how early the criterion is set out in Logic this is the likely source of this confusion over Popper’s demarcation criterion. Over time, I conjecture, little attention to the source material lead to the perpetuation of the myth that Popper's criterion is directed towards setting out the limits of the natural sciences.

Naturally, these problems are nonexistent in the German edition (1934/35). (In fact, I could only place variations of this formulation of the objection from ill-fit after 1959, the publication of the first English translation of Logik, and could find no instances of this objection from ill-fit in any German texts from 1934/35 to the present.)

(I can find no likely candidate passages in Logic for the popular myths that Popper's criterion is specifically about individual theories rather than theoretical systems or that Popper's territorial criterion is normative. In fact, Popper is explicit on the differences between normative and territorial problems and, as mentioned previously, the differences between theories and theoretical systems. I can only conjecture these myths began through secondhand transmission: it is likely that early exegesis on Logic was not done with proper care.)

7. Conclusion

Dismissing Popper’s territorial problem based on the previously formulated objection from ill-fit and characterising Popper’s stated territorial problem and criterion as such are both predicated on a mistake in exegesis. The objection from ill-fit confuses the problem as normative when it is territorial; as concerned with a scientific theory and not empirical theoretical systems; as setting out the differences between science and non-science when it is about setting out the absolute limits to empirical inquiry. This confusion is most likely due to Popper’s refusal to immediately flag the technical definition 'empirical science' in the 1959 English translation.

In summation, philosophers should stop perpetuating the objection from ill-fit presented previously as if it targets Popper’s territorial demarcation criterion. It may be effective against proposed territorial criteria that demarcate science from non-science; however, this objection is used repeatedly to target Popper’s criterion, and I can find no independent criteria that are not attributed to Popper.

There may be other reasons to reject any territorial demarcation criteria between empirical and non-empirical theoretical systems; this objection from ill-fit is not one of them: it is much ado about nothing, a shot aimed at an intellectual ghost.

Other objections from ill-fit directed against Popper’s criterion may succeed: while Church's objection does not target Popper's criterion, there remains the third version of the objection from ill-fit: Popper, Ayer and Carnap's criteria exclude universal probabilistic theories.


Ayer, A.J. (1936). Language, Truth and Logic. London: Victor Gollanz, 1st edition.

______. (1946). Language, Truth and Logic. London: Victor Gollanz, 2nd edition.

Boudry, M. (2013). Loki’s Wager and Laudan’s Error. In Boudry and Pigliucci (eds.) 2013: 79-98.

Boudry, M. and M. Pigliucci (eds.) (2013) Philosophy of Pseudoscience: Reconsidering the Demarcation Problem. Chicago: Chicago University Press.

Bunge, M. (1983). Demarcating Science from Pseudoscience, Fundamentia Scientiae 3: 369-388.

Carnap, R. (1937). Testability and Meaning--Continued. Philosophy of Science, 4(1): 1-40.

______. (1956). The Methodological Character of Theoretical Concepts. In The Foundations of Science and the Concepts of Psychology and Psychoanalysis. Edited by Feigl and Scriven. Minneapolis: University of Minneapolis Press: 38–76.

Church, A. (1949). Review of Ayer’s Language, Truth, and Logic, 2nd edition. The Journal of Symbolic Logic, 14(1):52–53.

Hempel, C. (1950). Problems and Changes in the Empiricist Criterion of Meaning, Revue Internationale de Philosophie, 41(11): 41-63.

______. (1951). The Concept of Cognitive Significance: A Reconsideration, Proceedings of the American Academy of Arts and Sciences, 80(1): 61-77.

Kitcher, P. (1982). Abusing Science. The Case Against Creationism, Cambridge, MA: MIT Press.

Lakatos, I. (1981). The Methodology of Scientific Research Programmes, Vol. 1, John Worroll and Gregory Currie (eds.), Cambridge: Cambridge University Press.

Laudan, L. (1983). The Demise of the Demarcation Problem. In R.S. Cohan and L. Laudan (eds.) Physics, Philosophy and Psychoanalysis, Dordrecht: Reidel: 111-127.

Nickles, T. (2013). The Problem of Demarcation: History and Future. In Boudry and Pigliucci (eds.) 2013: 101-120.

Mahner, M. (2013). How to Demarcate after the (Alleged) Demise of the Demarcation Problem. In Boudry and Pigliucci (eds.) 2013: 29-43.

Pigliucci, M. (2013). The Demarcation Problem. A (Belated) Response to Laudan. In Boudry and Pigliucci (eds.) 2013: 9-28.

Popper, K.R. (1934/5). Logik der Forschung: Zur Erkenntnistheorie der modernen Naturwissenschaft. Vol. 9 of Schriften zur wissenschaftlichen Weltauffassung. Vienna: Verlag von Julius Springer.

______. (1959. The Logic of Scientific Discovery, London: Hutchinson & Co. Translated and extended from the German original (Popper, 1935) by Karl Popper, Julius Freed, and Lan Freed. References are to Popper, K.R. (2000). The Logic of Scientific Discovery.

______. (1963). The demarcation between science and metaphysics. In The Philosophy of Rudolf Carnap. Vol. XI in Library of Living Philosophers. Edited by P.A. Schilpp. La Salle: Open Court Press: 183–226.

______. (1983). Postscript to the Logic of Scientific Discovery: Realism and the Aim of Science, vol. iii. Originally circulated as private galley proofs in 1956/1957.

Preston, J. (1994). Methodology, Epistemology and Conventions: Popper’s Bad Start, PSA, 1994(1): 314-322.

Putnam, H. (1974). The Corroboration of Theories. In Schilpp (ed.) 1974. The Philosophy of Karl Popper, The Library of Living Philosophers, vol xiv, books i-ii. La Salle: Open Court Press, book i.

Thagard, P. (1978). Why Astrology is a Pseudoscience, PSA, 1978(1): 223-234.

Thornton, S. (2016). Karl Popper, The Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/win2016/entries/popper/>.

Waibl, E. and P. Herdina (1997). German Dictionary of Philosophical Terms, Vol 2: English-German, London: Routledge.


No comments:

Post a Comment

Template developed by Confluent Forms LLC; more resources at BlogXpertise