Scientism and the Problem of Self-Referential Incoherence

ABSTRACT

In recent discussions on the possibilities and boundaries of science, the term ‘scientism’ has been used with multiple meanings. In this paper, this term is taken to refer to a strong epistemic thesis according to which scientific research is the only reliable source of knowledge. Strong epistemic scientism has been recently criticized from different points of view. A recurrent accusation is that the thesis is self-referentially incoherent. After giving an overview of the critical arguments and the answers given to them, I will try to point out that the debate concerning the correctness of scientism is based on a questionable background assumption about the role of higher-order justification.

KEYWORDS:

• Epistemic scientism
• self-reference
• incoherence
• first-order theory vs. higher-order theory

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. On my view, Robinson (2015) represents a clear example for this interpretive situation.

2. For a similar definition, see de Ridder (2014, 25).

3. An early example for this narrowly understood scientific epistemology is Mahner and Bunge (1996). They claim that ‘neither art, nor music, nor literature make any factual claims as to the nature and history of the world’ (Mahner and Bunge 1996, 193).

4. For an objection of this kind, see Plantinga (2018). A referee raised a similar objection: ‘The idea that mathematics is not part of natural sciences seems puzzling. How could we have had the advancements in science without mathematics?’ The best way to answer this objection is to make a distinction between theoretical mathematics and applied mathematics. Since theoretical mathematics is an a priori discipline, it does not count as a natural science. But the methods of applied mathematics are compatible with the methods of experimental natural sciences.

5. Ross also acknowledges the social sciences as sciences (Ross 2017, 227) and arguably, so does Ladyman. In addition, Ladyman accepts that the humanities contribute to objective knowledge (Ladyman 2018, 106n1, 107–108, 123–124).

6. Hirvonen and Karisto (2022) provides a useful overview of the current standing of the debate on the demarcation problem. As a referee has helpfully pointed out, there are some works where the issues of scientism are treated in parallel with the demarcation problem. See, for instance, Ladyman and Ross (2007, 7, 28, 33–34) and Turunen et al. (2023, Sect. 3).

7. It should be noted here that Stenmark rejects SCIENTISM because of its self-referential character. We can reconstruct his argument as follows (see, Stenmark 2001, 32): We cannot come to know SCIENTISM by appeal to science alone. It is rather a view in the theory of knowledge and is, therefore, a piece of philosophy and not a piece of science. But if this is the case, then SCIENTISM is self-refuting. If SCIENTISM is true, then it is false. SCIENTISM falsifies itself. For more on this, see also Stenmark (2018).

8. Barker and Kitcher (2014, 74–76) are an exception in this regard. One might think that self-reference yields circularity in the justification of scientism. But Barker and Kitcher argues that justification follows in fact a spiral line: scientists revise their methods continuously and can therefore invent better and better justificatory procedures. Unfortunately, pursuing this proposal has to be deferred to another occasion.

9. A similar position is to be found in Buckwalter and Turri (2018, 281). They call their approach to science moderate scientism. According to this approach ‘science can help answer questions in disciplines typically thought to fall outside of science’.

10. As a referee remarked, it’s hard to see how the issue of the quantity of publications is even relevant. Mizrahi (2017, 356) defends his point by saying that ‘something can be better than something else either quantitatively or qualitatively’. This argument is too general to be acceptable in this context, so I agree with the referee’s remark. For the sake of argument, however, I assume here that Mizrahi is right.

11. Concerning this point, a referee offers a counter argument: the basic argument for scientism is that science is more successful than all other epistemic projects. So, what we only need is agreement on the success criteria. If all or most grounds for comparison (say, science, common sense, etc.) agree on the criteria for success and failure, then there isn’t any problem for WEAK SCIENTISM. If there are no commonly shared criteria for success, then we have to endorse epistemic relativism. So the justification of WEAK SCIENTISM is something that can be determined. How can one respond to this objection? I think there are no commonly shared criteria for success. But it seems to me that even if we are willing to endorse some kind of epistemic relativism, there still remains an internal problem for WEAK SCIENTISM: if someone insists that scientific research is the best source of knowledge, then her position presupposes the availability of an anti-relativist frame of success evaluation. Something can be said to be the best source of knowledge only if there is a criterion that can uniformly be applied across all relevant epistemic domains. This leads us back to the original problem. For more on this argumentative situation, see Baghramian and Carter (2022).

12. To be honest, Hietanen et al. (2020, 541) recognize this difficulty by saying that ‘[e]ven if scientism could not be accepted as a scientific hypothesis (which, of course, is not the case), it could still be validated by using non-scientific means. In this case, scientism would not be justified in the best possible manner, but it could be justified nevertheless, in the same sense that our everyday judgements can be justified’. However, as the quoted sentences plainly show, they regard this difficulty as a manageable one.

13. Concerning this point, a referee critically remarks that there are truths that are unknown. Metamathematics is not unified concerning the interpretation of mathematical truth (cf. Linnebo 2017). Some think that truth coincides with provability in an axiomatic system. This view entails that all mathematical truths are known. Others think that mathematical truths exist independently from human knowledge. So there may be mathematical truths that are not known. In this sense, the referee’s remark is correct. I do not want to delve into this complicated topic, however, so I restrict myself to referring to Burgess’ paper.

14. According to mathematical conventionalism, mathematical knowledge is fully justified by the linguistic conventions of mathematical language. For a novel defense of this old view, see Warren (2020).

15. One might ask: if the epistemically significant work is performed always at the first-order level of research, then how are paradigm changes in science possible? This is an important question. When a scientific theory turns out to be false and there seems to be no first-order way out from the situation, higher-order or metatheoretical reflection may offer a solution. In cases like this, even a priori conceptual analysis may play a partial role in empirical theory revision. The examination of this complex process is not directly relevant to the scientism debate, however.

16. I would like to thank three referees for valuable comments and suggestions improving the clarity of my arguments.

Additional information

Funding

This work was supported by the Subprogramme for Linguistic Identification of Fake News and Pseudo-Scientific Views, part of the Science for the Hungarian Language National Programme of the Hungarian Academy of Sciences (MTA).

Notes on contributors

Zoltán Vecsey

Zoltán Vecsey is a research fellow at the MTA-SZTE-DE Research Group for Theoretical Linguistics and Informatics. His research interests include the foundational questions of scientific realism, the problems of nominal semantics and the theory of fictional discourse.