Philosophy of Science 101: Anti-Realism

Foreword

The Philosophy of Science series explores both general questions about the nature of science and specific foundational issues related to the individual sciences. When applied to such subject areas, philosophy is particularly good at illuminating our general understanding of the sciences. This 101 series will investigate what kinds of serious—often unanswered—questions a philosophical approach to science exposes through its heuristic lens. This series, more specifically, will look at the ‘Scientific Realism’ debate throughout, which questions the very content of our best scientific theories and models.

Philosophy of Science 101 will be divided into the following chapters of content:

1. Philosophy of Science 101: The Relationship Between Philosophy and Science

2. Philosophy of Science 101: Scientific Realism

3. Philosophy of Science 101: Anti-Realism

4. Philosophy of Science 101: Realism and Anti-Realism ‘Compromise’

5. Philosophy of Science 101: Causation

6. Philosophy of Science 101: Scientific Models

7. Philosophy of Science 101: Models of Explanation

8. Philosophy of Science 101: Laws of Nature

9. Philosophy of Science 101: Science and Social Context

Philosophy of Science 101: Anti-Realism

Previously, this 101 series introduced the scientific realism debate. The debate is a particularly significant part of the philosophy of science enterprise which explores scientific "knowledge" in great depth (Chakravartty, 2017), considering the ampliative nature of scientific theories that often go far beyond what is directly observable via human sensory capabilities. Scientific realism, on the one hand, provides a positive epistemic attitude toward the content of the best theories and models in science. Realism indeed recommends belief in both observable and unobservable aspects of the world described by the sciences, as was the primary focus of the preceding 101 article.

Each dimension of scientific realism discussed last time—namely metaphysical, epistemic, and semantic—may progressively defend itself against varying "levels" of the rival epistemologies of science, known collectively as forms of scientific anti-realism (recall how this is in the debate over the use of unobservable entities and mechanisms by scientific theories to describe phenomena that is observable) (Psillos, 1999). Having considered the scientific realist position on this debate, this third article in the series will now investigate the said rival epistemologies of science: anti-realism.

Figure 1: David Hume's empiricism is at the very root of the scientific realism debate (Jessop, 2022).

Recap: Hilary Putnam’s "No Miracles" Argument

Before delving into anti-realism, it is worth taking a brief look back at scientific realism first since each anti-realist position results from the given realist dimension it rejects. Consider the famous "no miracles" argument presented by Hilary Putnam (1975) in favour of scientific realism. Putnam’s argument is possibly the most influential realist argument (Dawid & Hartmann, 2018), suggesting that scientific realism can provide an explanation of the success of science. According to Putnam (1975), science often gets things right. The argument notably begins with the widely accepted premise that the best scientific theories are extraordinarily successful because they facilitate empirical predictions, retrodictions, and explanations of the subject matters of scientific investigation, often marked by astounding accuracy and intricate causal manipulations of the relevant phenomena. What might explain this success then? An explanation favoured by scientific realists—largely due to Putnam’s contribution—is that the best theories are in fact (approximately) true, which is to say that they correctly describe a mind-independent world (Dawid & Hartmann, 2018). Hence there are no miracles according to Putnam. Miracles simply may not account for the success of science, and if successful theories were far from the truth, their vast success would be entirely miraculous. The best, most successful, scientific theories give true (or nearly true) descriptions of observable and unobservable aspects of a mind-independent world on this account. Inference to the best explanation (or "IBE") may also support Putnam’s argument here, since scientific realism supposedly proves better than any other philosophical position to explain the success of science. So the argument goes. This article will next introduce various opposing anti-realist stances, ones which object to the very idea that realism “is the only philosophy that doesn’t make the success of science a miracle”, as Putnam (1975) puts it (p. 73).

Anti-realism: Reductive Empiricism

As mentioned, there are numerous different "brands" of anti-realism depending on how many realist theses one denies. Take an example: reductive empiricism. This particular anti-realist view denies two particular premises of realism: (a) semantic realism, that scientific theories provide literal descriptions of their subject matter and may be true or false where terms in a theory refer to the entities they postulate, and (b) epistemic optimism that mature and predictively successful scientific theories are true of the world and humans can know them to be so (Psillos, 1999). Since the reductive empiricist denies such realist premises, it is argued that theoretical language (i.e., scientific language and terms about unobservables) (Chakravartty, 2017) should not be taken literally. As such, one must "reduce" talk about theoretical entities to talk about observable entities (Feyerabend, 1962). In this sense, theoretical terms are kind of like metaphors instead of literal expressions. The leading idea being that one must understand theoretical claims as disguised claims about observable entities. The so-called ‘languages’ may therefore be separated into two categories, namely (a) observation language, and (b) theoretical language (Curd & Cover, 1998). Here, theoretical terms in scientific theories can be given explicit definitions in (a). Simple examples of observation language include "round", "red", "liquid", or "ball" (i.e., the kinds of properties that one can observe). Examples of theoretical language, on the other hand, include "proton", "magnetic field", or perhaps "inflation". To explain the (anti-realist) reductive empiricist view further, consider the following (with a theoretical term T and observational term O):

For all x: Tx ↔ Ox

This is to say that for every theoretical term (T), there must also be an observational term (O). For instance, for all x: x has a temperature T iff (if and only if) a mercury thermometer that is in contact with x shows an observable outcome O. Hence reduction.

Figure 2: Photograph of the late Bertrand Russell. Reductive empiricism cannot apply to Russell's logical formalism (Monk, 2022).

Indeed, an explicit definition is given to a theoretical term in reductive empiricism (Curd & Cover, 1998). To reiterate, since reductive empiricism denies semantic realism, one must "reduce" talk about theoretical entities to talk about observables. Theoretical terms can therefore be given explicit definitions via observation language. Unlike what the realist view suggests, it is simply not true that scientific theories may provide literal descriptions of their subject matter (Curd & Cover, 1998). A few problems arise with reductive empiricism here, however. Firstly, the meaning of theoretical concepts is not exhausted by observations. Moreover, every new method of observation gives rise to a different concept. So, one ends up with a multiplicity of concepts where, intuitively, there could only be one concept. Stathis Psillos (1999) describes just how problematic reductive empiricism is, starting with the reductive empiricist move to reduce the content of the world to whatever is observable (hence the aim to translate talk employing theoretical terms into talk employing only observational terms and predicates). The problem with this move is its patent failure, says Psillos (1999). It is now generally agreed that that theoretical concepts have excess content, that their meanings are fixed by the theories in which they are embedded and that they have putative factual reference. In this setting, the irreducible existence of unobservable entities is enough for their independent existence (Psillos, 1999). Though this article will not consider such problems in any significant detail, it is worth noting that one might give up reductive empiricism because of them. One might instead adopt constructive empiricism (van Fraassen, 1980).

Anti-realism: Constructive Empiricism

Coined by Bas van Fraassen (1980), constructive empiricism is another form of anti-realism which—unlike reductive empiricism—grants both metaphysical and semantic realism but denies epistemic optimism (the realist view that mature and predictively successful scientific theories are true of the world and that humans can know them to be so). Put differently, constructive empiricism holds that (a) the world exists, and (b) the world has a mind-independent structure. Further, the statements made by a scientific theory about unobservables should also be taken literally. They are indeed true or false depending on whether the entities they describe are part of the mind-independent world and on whether these entities behave in the way that the theory predicts (van Fraassen, 1980). Accepting the best theories science has to offer, however, does not require belief in the unobservable entities postulated by these theories. Unlike Putnam’s (1975) "no miracles" argument described earlier, the success of such theories can instead be understood without invoking the existence of unobservable entities (Rosen, 1994). On this view, science aims to provide theories which are merely empirically adequate. According to van Fraassen (1980), the constructive empiricist therefore disagrees with the realist about the aim of science. Whereas scientific realists take science to aim at full-fledged truth, constructive empiricists hold that science only aims to tell the truth about what is observable (van Fraassen, 1980).

Figure 3: A Theory’s Empirical Adequacy, part of constructive empiricism (Monton, 2007).

Arguments For Anti-realism: Empiricism and Underdetermination of Theory by Evidence

Regardless of how convincing or far-fetched particular accounts of anti-realism like van Fraassen’s might prove, it is important to consider why one might adopt anti-realism in any of its forms (instead of scientific realism) in the first place—starting with empiricism. Empiricism is the view that sense experience is the only source of knowledge (Sellars, Rorty, & Brandom, 1997). In its strictest sense, empiricism is an important motivator behind anti-realism since it holds that there is simply no epistemic warrant for anything over and above what is given in direct experience. This means that there is no warrant for unobservable entities postulated by scientific theories, despite their ampliative nature. If such entities that are postulated in scientific theories are unobservable, empiricism says that one cannot gain any knowledge of them—or their existence—as such. Unlike what realism might suggest about science, or rationalism (i.e., the view that humans can intuitively gain knowledge), it is simply not possible on the empiricist view that humans can have knowledge of something which they cannot observe. Knowledge of an entity and its existence may instead come predominantly from experiences gathered through the five senses (Sellars, Rorty, & Brandom, 1997). In empiricism, knowledge is spoken of as a posteriori, or from the latter, meaning gained from experience (Sellars, Rorty, & Brandom, 1997). Hence one might reject scientific realism and adopt anti-realism for this very reason, denying that scientific theories can provide any true statements about unobservable entities in a mind-independent world. The success of scientific theories may not be explained by the truth of statements about unobservables (Putnam, 1975) in empiricism—humans may not gain knowledge of something they cannot observe.

Now consider underdetermination, another argument for anti-realism. The "underdetermination of theory by evidence" argument points out that often more than one theory is compatible with the available evidence. Therefore, on the basis of evidence alone, one cannot decide which of the alternatives (if any) is true (Bonk, 2008). The general argument is as follows:

Premise 1). The evidence available is E.

Premise 2). Theory T is consistent with E.

Premise 3). Theory T’ is consistent with E.

Premise 4). T and T’ are not equivalent.

Conclusion. There is no reason to take T rather than T’ to be true (or vice versa).

Figure 4: Underdetermination: plotted points (data), yield unimaginably many curves (theories) (Norton, 2009).

When presented on a graph, this might also be known as curve fitting. Two (or many) different theories are compatible with the evidence, but not with each other. The history of science indeed provides infamous examples of this; wave and particle optics in the 18th century, for instance, or even more contemporary theories like superstring or quantum (Butterfield, 2014). The issue is that, if one cannot decide which of the several non-equivalent alternatives is true, then how can one ever be realist if it is impossible to know which theory to be realist about?

Putnam’s (1975) "no miracles" argument is seemingly supported by IBE, as discussed at length in the previous 101 article. The underdetermination of theory by evidence, however, presents a serious challenge to IBE and scientific realism here. As an IBE-based argument, Putnam’s (1975) "no miracles" holds that scientific realism proves better than any other philosophical position to explain the success of science. The governing idea—via IBE—is that explanatory considerations are a guide to inference (i.e., a conclusion reached on the basis of evidence and reasoning), that scientists infer from the available evidence to the hypothesis which would, if correct, best explain that evidence (Lipton, 2017). In the context of Putnam’s argument, "evidence" notably refers to the vast success of scientific theories. But what about when the same evidence support when multiple theories, incompatible with one another? Underdetermination of theory by evidence appears to undermine the IBE logic here, and one must question whether an argument like Putnam’s can genuinely be supported by IBE—or taken seriously—if available evidence at a given time is insufficient to determine what beliefs should be held in response to it (Stanford, 2009).

Figure 5: American Philosopher Grover Maxwell (1962) suggests a continuum of observable and theoretical entities.

Science Has a History

Scientific realism suggests that mature scientific theories have to be successful, and it affords us an explanation of why this is so. According to the no miracles argument, realism explains the success of scientific theories better than any other philosophical account (Putnam, 1975). This is based on the evidence that mature scientific theories are successful. By IBE, then, scientific realism is true. This article has already considered one challenge to IBE, namely underdetermination. Importantly, another challenge (somewhat similarly) comes from the history of science. According to many, scientific realism is highly implausible because of this history. Science is in fact full of theories that were once empirically successful and yet turned out to be false (de Paz, 2021). This argument supporting anti-realism is known as pessimistic meta-induction (PMI), since via meta-induction, the empirical success of science can be shown not to warrant the claim that theories are (approximately) true (Doppelt, 2007). Consider a few examples from the history of science to reflect this pessimistic argument, first the shift from the Ptolemaic system (the once accepted scientific theory that the earth is at the centre of the universe) to the Copernican system (i.e., the sun at the centre of the universe). Second, the movement from Newtonian physics to the theory of relativity and to quantum physics (Kuhn, 1970). The PMI is indeed a direct response to Putnam’s no miracles argument and is closely tied to Thomas Kuhn’s (1970) well-known theory of scientific revolutions, all of which shall be discussed in the next article of this series (including the compatibility of Kuhn’s theory with scientific realism, and a potential so-called "compromise" between realism and anti-realism).

Conclusion

The basic idea of realism is that the kinds of thing which exist, and what they are like, are independent of us and in the way in which we find out about them; anti-realism denies this (Craig, 1998). Anti-realism is in fact defined in opposition to realism, and so it is natural to ask first what realism is so to arrive at a characterisation of anti-realism on this basis (Rosenkranz, 2013). This article has done exactly that, starting with a brief recap on the preceding 101 article, in particular on Hilary Putnam’s (1975) prominent no miracles argument in favour of scientific realism. It is important to note that each form or account of anti-realism depends on the various parts of realism it denies – this article therefore started by outlining reductive empiricism, the kind of anti-realism which reduces talk about theoretical entities to talk about observables (thus denying semantic realism and epistemic optimism as a result). There are various problems with this form of anti-realism, however, concerning the separate observation and theoretical languages. One might therefore give up reductive empiricism in favour of van Fraassen’s (1980) constructive empiricism, which only denies one specific part of scientific realism that requires belief in unobservable entities postulated by scientific theories. Both reductive and constructive empiricism provide good examples of the different ways one might adopt anti-realism. This article finally considered the broader arguments for anti-realism: empiricism and underdetermination of theory by evidence. The named arguments do not point to a particular type of anti-realism, but they do oppose specific parts of scientific realism (like IBE). Underdetermination in particular goes against IBE and the no miracles argument, suggesting that evidence alone is not enough. The history of science indeed ties in nicely with this and plays an important role in the scientific realism debate since currently successful theories will very probably turn out to be false too. The history of science and Kuhn’s (1970) argument on scientific revolutions will therefore be explored in the next 101 article of the series.

Bibliographical References

Bonk, T. (2008). Underdetermination: An Essay on Evidence and the Limits of Natural Knowledge, Dordrecht, The Netherlands: Springer.

Butterfield, J. (2014). “On Underdetermination in Cosmology”, Studies in History and Philosophy of Modern Physics, 46: 57–69.

Chakravartty, A. (2017). Scientific Realism. In The {Stanford} Encyclopedia of Philosophy (Summer 2017). Metaphysics Research Lab, Stanford University.

Craig, E. (1998). Realism and antirealism. In The Routledge Encyclopedia of Philosophy. Taylor and Francis. Retrieved 13 Jan. 2023, from https://www.rep.routledge.com/articles/thematic/realism-and-antirealism/v-1. doi:10.4324/9780415249126-N049-1

Curd, M., & Cover, J. A. (1998). Philosophy of science: The central issues.

Dawid, R., & Hartmann, S. (2018). The no miracles argument without the base rate fallacy. Synthese 195, 4063–4079. https://doi.org/10.1007/s11229-017-1408-x

de Paz, M. (2021). “Poincaré, Le Roy, and the Nouveau positivisme”, HOPOS: The Journal of the International Society for the History of Philosophy of Science, 11(2): 446–460. doi.org:10.1086/715880.

Doppelt, G. (2007). Reconstructing scientific realism to rebut the pessimistic meta-induction. Philosophy of Science, 74(1), 96-118.

Feyerabend, P. K. (1962). Explanation, reduction, and empiricism.

Kuhn, T. S. (1970). The structure of scientific revolutions (Vol. 111). University of Chicago Press: Chicago.

Lipton, P. (2017). Inference to the best explanation. A Companion to the Philosophy of Science, 184–193. https://doi.org/10.1002/9781405164481.ch29

Putnam, H. (1975). Mathematics, Matter and Method, Cambridge: Cambridge University Press.

Psillos, S. (1999). Scientific Realism: How Science Tracks Truth, London: Routledge.

Rosen, G. (1994). What is constructive empiricism?. Philosophical studies, 74(2), 143-178.

Rosenkranz, S. (2013). Realism and Anti-Realism. obo in Philosophy. doi: 10.1093/obo/9780195396577-0098

Sellars, W., Rorty, R., & Brandom, R. (1997). Empiricism and the Philosophy of Mind (Vol. 1). Harvard University Press.

Stanford, K. (2009). “Underdetermination of Scientific Theory”, The Stanford Encyclopedia of Philosophy (Winter 2009 Edition), Edward N. Zalta (ed.), URL=<https://plato.stanford.edu/archives/win2009/entries/scientific-underdetermination/>

van Fraassen, Bas C., (1980), The Scientific Image, Oxford: Oxford University Press.

Visual Sources

Cover Image: Pils, R. (2020). Quine's Scientific Realism Revisited. Theoria, 86(5), 612-642.

Figure 1: Jessop, T. E. & Cranston, M. (2022, October 28). David Hume. Encyclopedia Britannica. https://www.britannica.com/biography/David-Hume

Figure 2: Monk, R. (2022, December 5). Bertrand Russell. Encyclopedia Britannica. https://www.britannica.com/biography/Bertrand-Russell

Figure 3: Monton, B. (ed. 2007). Images of Empiricism: Essays on Science and Stances, with a Reply from Bas C. van Fraassen, Oxford: Oxford University Press.

Figure 4: Norton, J.D. (2009) Underdetermination, Center for Philosophy of Science. Available at: https://sites.pitt.edu/~pittcntr/Being_here/last_donut/donut_2008-09/underdetermination_03-21-09.htm (Accessed: January 16, 2023).

Figure 5: Maxwell, Grover (1962). The ontological status of theoretical entities. In Herbert Feigl & Grover Maxwell (eds.), Scientific Explanation, Space, and Time: Minnesota Studies in the Philosophy of Science. University of Minnesota Press. pp. 181-192.