In his review of Hasok Chang’s new book (Chang 2023) in this Journal, Daniel Brooks focuses on its visionary aspects, like the “opportunity for those open to or seeking reorientation in how we analyze processes of scientific knowledge production”, and Chang’s attempt at “cultivating a new spirit of discourse [about the issues of philosophy of science] among its participants” (Brooks 2025, 143). In this discussion note, we want to complement Brooks’ insightful review by a critical discussion of some more concrete topics, much discussed in the philosophy of science, that Chang engages with, like correspondence, truth, realism and pluralism. Before doing so, let us first describe our take on Chang’s project in the book. In our view, Chang mainly defends two positions. The first is a new form of pragmatism, as can be seen from the subtitle of the book: “A New Pragmatist Philosophy of Science”, or from section titles like “Pragmatist Notions of Knowledge, Truth and Reality” (3) or “Rehabilitating the Pragmatists” (197).^Footnote1 The scope of his pragmatist ambition is very wide indeed, because it even covers logic (65–67): “It would have been the ultimate prize for a pragmatist to argue successfully that even logic was only pragmatically justified” (65). The second claim Chang defends in this book is pluralism. The scope of his pluralist ambition is also very wide because he defends many kinds of pluralism—conceptual, epistemic, ontological, concerning progress and truth. As Chang states, “pluralism is a key aspect of the whole outlook on knowledge that I am advocating in this book” (131). We may note without further discussion, that Chang’s engagement with pluralism, combined with his explicitly humanist stance (62, 63, 252), make him the most important contemporary successor of Paul Feyerabend (see also Chang’s references to Feyerabend on pages 102, 215, 232, 236, 238).

We shall proceed as follows. In Sect. 2, we shall discuss Chang’s criticism of the correspondence theory of truth and of what he calls ‘correspondence realism’. Section 3 takes up Chang’s alternative to the correspondence theory of truth, his ‘truth-by-operational-coherence’. The topics of both Sects. 2 and 3 are strongly connected with the questions of realism and anti-realism, and we will discuss Chang’s postulated realism in Sect. 4. Before concluding our review, in Sect. 5 we discuss the tension that exists in Chang’s book between his vigorous defence of pluralism on the object-level and his apparently taken-for-granted monism on the metalevel, namely pragmatism.

Correspondence is an important topic in the book, especially the critique of what Chang calls ‘correspondence realism’, which means that “the truth of theories consists of their correspondence to the mind-independent world” (69). This critique strategically prepares the ground for his defence of an important pillar of pragmatism, a pragmatist theory of truth (see Sect. 3). Chang makes a very useful distinction regarding mind-independence. He distinguishes between “mind-control” and “mind-framing” (71). Mind-independence in the sense of lacking mind-control means that “our minds do not have direct wilful control over any reality outside of ourselves” (71), which is a core realist intuition that Chang fully shares. Quite another thing is mind-framing, which Chang thinks is unavoidable: “real entities are mind-framed because all entities are mind-framed. All that we can ever think or talk about are mind-framed entities” (71). Chang supposes that many debates about realism suffer from the “ambiguous notion of mind-independence” (73) by identifying it with mind-control and thereby overlooking the possibility and necessity of mind-framing. This is also the main objection against what Chang calls ‘correspondence realism’: the idea that “the truth of theories consists in their correspondence to the mind-independent world” (69). Although Chang defends pluralism about truth (167), he rejects the correspondence theory of truth as a possibility of “primary” truth. A proposition is “true in the primary sense” if its “truth does not derive from the truth of other propositions” (166). Since the correspondence theory of truth is a “non-operational idea” (164), because we have no access to mind-independent facts,^Footnote2 it must be given up in the domain of empirical phenomena and replaced by something else. Chang proposes his version of a pragmatist theory of truth, called “truth-by-operational-coherence” (167). We, however, propose a different therapy for the standard conception of the correspondence theory of truth.

The standard view in analytic philosophy of the correspondence theory of truth (CT) claims a correspondence of propositions (or theories etc.) with mind-independent facts or the mind-independent world.^Footnote3 This theory merges two very heterogenous elements into one ‘definition’ (or ‘theory’): the correspondence idea (a relation) on the one hand and metaphysical realism, a specific assumption about one of the relata, on the other. This is a highly artificial philosophical construction that does not always correspond with our pre-philosophical use of ‘truth’, of which it should be an explication. CT seems to capture the truth of the following example well: ‘Mount Everest is higher than the Matterhorn’.^Footnote4 Indeed, the fact of Mount-Everest-being higher-than-the-Matterhorn seems to be mind-independent and CT seems to be unproblematically applicable. This is different in our second example. Suppose the following dialogue. Question: Did you cheat on me yesterday? Answer: No, I did not cheat on you yesterday. Clearly, the truth of the answer depends on whether s/he cheated yesterday or not, that is, on this fact. This fact, however, is not mind-independent because ‘cheating’ is culture- and may even be couple-dependent: one and the same physical act may be cheating for one couple and something below cheating for another. Here, it becomes obvious that Chang’s distinction between mind-control and mind-framing is extremely useful (71). Whether cheating took place in a particular situation in the past is mind-framed, but not mind-controlled. Thus, in everyday discourse, the idea of truth as a correspondence between propositions (etc.) and (mind-framed) facts is entirely unproblematic (with which Chang agrees, 79). It is obvious that this idea of truth is of utmost importance in our non-philosophical life, not only for couples regarding their faithfulness, but also for courts, politics, the economy, etc. It is difficult to see how in these discourses the element that is constitutive of the understanding of truth, namely correspondence, could or should be given up.

This also holds for objects of our everyday world that are mind-framed in an even deeper way, namely, coloured objects, assuming that colours are secondary qualities. In many cases it is completely unproblematic to understand what it means for a sentence like ‘this is a red sweater’ to be true or not and to decide its truth. However, truth as correspondence can also be used in a context in which talk of colours is not legitimate any longer, say in an investigation in physical optics in which colours do not exist, but only spectral distributions, reflective surfaces, etc. We can go even further and note that also a Kantian can use truth as correspondence in a world in which all physical things are mind-framed by the forms of intuition and the categories.^Footnote5 However, a metaphysical realist may also use truth as correspondence, referring to a completely mind-independent world.

We can glean from these examples that a properly understood correspondence theory of truth is metaphysically neutral once it is liberated from the forced and unhappy marriage with metaphysical realism. Truth as correspondence can be coupled with many different ontologies: physical things, Kantian appearances, Kuhnian paradigm-shaped worlds, tropes, processes, Wittgensteinian facts, Platonic ideas, and also metaphysical realism.^Footnote6 The freedom of choice regarding ontologies goes so far that also undoubtedly counter-factual ontologies can be coupled with the correspondence idea of truth, for instance Harry Potter’s or Sherlock Holmes’ universes. It makes perfect sense to claim that it is true that Harry Potter went to Hogwarts School of Witchcraft and Wizardry. The notion of truth in this claim is the same as in its everyday uses, the ‘world’ it refers to is the world that is described in several volumes written by J. K. Rowling.

Thus, truth as correspondence when divorced from any necessary connection to metaphysical realism, is fully ‘operational’, that is, applicable, in many different contexts. Thus, we are by no means forced to give up the correspondence notion of truth once we have freed it from metaphysical realism and replace it by an alternative. However, even if we do not buy into Chang’s argument about the inapplicability of the correspondence theory of truth, we should examine Chang’s alternative conception of truth, ‘truth-by-operational-coherence’, because it may be superior to a correspondence notion, perhaps globally, perhaps only for certain domains.

As already mentioned, Chang claims to extend) his pluralism also to truth: “I subscribe to pluralism concerning truth: there are many different notions of truth, which have different uses in various domains and contexts” (167). However, this is not a fully democratic pluralism in which every member of the community has the same rights. One member clearly stands out, truth-by-operational-coherence: “But I do want to propose that truth-by-operational-coherence is what constitutes primary truth in empirical domains” (167). And these empirical domains include “science and much of daily life as well” (167). Note that this is again a very ambitious project, in at least two interrelated respects. First, Chang intentionally says, “constitutes primary truth” and does not say “indicates” primary truth, and he leaves no doubt, why: “The operational coherence of activities relying on a true proposition are not consequences or indications of its truth, from which we may infer the truth. Rather, operational coherence is constitutive of truth” (170; emphasis in the orginal). Expressed with the common distinction between a ‘definition’ of truth and ‘criteria’ of truth, Chang means a definition of truth, not an additional criterion of truth.^Footnote7 “I am proposing what we ought to mean by ‘truth’ in order to render it as a useful concept” (170; emphasis in the original). The motive for this strong claim is that Chang rejects the correspondence theory of truth in its standard understanding for primary, that is, non-derived (166), empirical truth (170; see above, Sect. 2). The envisaged replacement for the discarded correspondence theory of truth is truth-by-operational-coherence.

Also in a second respect, Chang’s claim about truth-by-operational-coherence is a strong claim: “truth-by-operational-coherence is what constitutes primary truth in empirical domains” (167; emphasis added). In other words, the other truth conceptions are, at least as far as empirical knowledge is concerned, secondary, derived, non-autonomous, because they are dependent on primary truth. Here is Chang’s characterization of truth-by-operational-coherence: “A proposition is true to the extent that there are operationally coherent activities that can be performed by relying on it” (167; emphasis in the original). Let us exemplify truth-by-operational-coherence by an example. Consider the following sentence: ‘A W 1/2″ bolt (British Standard Whitworth) will not fit a M12 nut (German Standard) but will fit a 1/2″ nut.’ Now I try to thread a W 1/2″ bolt first through a M12 nut and then through a 1/2″ nut. The first action consistently fails whereas the second consistently succeeds. It seems indeed that the whole meaning of the sentence and its truth is fully captured by the failing and the succeeding actions. There seems to be nothing ‘theoretical’ in the sentence over and above the actions. Also, the falsity of a sentence should mean that the respective action cannot be performed, as would be demonstrated by ‘A W 1/2″ bolt fits an M12 nut’—indeed, you cannot thread the bolt through that nut. All of this, of course, can motivate pragmatism, especially Chang’s version of it.

Because the realm of empirical phenomena for which truth-by-operational-coherence is supposed to be appropriate is not limited, let us try out another example. Suppose A, the partner of B, truthfully tells B: ‘I cheated on you yesterday’. What could the ‘operationally coherent activities’ be ‘that can be performed’ by B due to his or her relying on the truth of A’s statement ‘I cheated on you yesterday’? A possible reaction of B could be to leave A. But what would it mean if B finds themself unable to do that? Another possible reaction of B could be to treat A like sh*t for weeks and months to make A feel miserable and guilty. Would that really constitute the truth of ‘A cheated on B’? Thus, it is difficult to see to which actions specifically the truth of “I cheated on you yesterday” is tied. However, it is even harder to see the performance of which actions would have been impossible for B if A had not said the truth, that is, if A did in fact not cheat on B, but lied with the confession. None of B’s actions that reacted to A’s supposed cheating would have been impossible to perform if A had, in fact, not cheated.

It seems to us that there are many such cases in which the truth of a sentence is only very loosely coupled to actions, if at all. In the cheating case, for example, undoubtedly B’s respective actions refer to A’s reported cheating, but already to say that they express the truth of A’s report seems extremely far-fetched, let alone that they are constitutive of the report’s truth. Is it not much more plausible to say that the sentence ‘A cheated on B’ is true if and only if in that particular social world framed by A and B, A indeed cheated on B? Does this fact not exist, if it really happened, independently of all subsequent actions, and is the truth-maker of the sentence ‘A cheated on B’?

Here is a consequence of truth-by-operational-coherence that Chang clearly recognizes and endorses. We, however, find it a serious weakness of truth-by-operational-coherence because it flattens a distinction that is relevant in science and therefore relevant for philosophy of science. The distinction concerns successful, intentionally ontologically false models versus successful approximately true models (185–186; emphasis in the original). Note that this distinction by itself in no way implies any of the more robust forms of realism, because it is made in the respective mind-framed world. Here is an example: In the late 1920s, physicists began to invent the liquid drop model of the atomic nucleus. It was known shortly after that nuclei mainly consisted of protons and neutrons, the most common isotope of uranium ²³⁸U, for instance, of 92 protons and 146 neutrons. This many-body problem was at the time intractable as such, so physicists counterfactually modeled the nucleus as a liquid drop, that is as a continuous fluid, with viscosity, surface tension and electric charge.^Footnote8 This was an immensely successful model: it served, among other things, as the basis for the development of atomic bombs with different designs. However, everyone knew that this model was massively counterfactual, a mere analogy, that is, ontologically false, because the mind-framed world contained particles as the fundamental constituents of atomic nuclei.^Footnote9 It may make a difference for the research strategy of a scientist whether she conceives of some assumption in the model as counterfactual and purely analogical, or as representative of reality. In the former case, she will have less inhibitions to modify it in various respects in order to see what the consequences are, than in the latter case. Thus, in order to understand scientific behavior, we must know whether the scientists involved conceive of a model as approximately realist or as counterfactual. For Chang, however, as only operational success counts for truth, he calls such successful counterfactual models “true-by-operational-coherence at least in a limited domain” (186). We, however, fail to see the advantage of assimilating these two sometimes very different kinds of models by calling both of them true.

Thus, we do not find Chang’s version of pragmatist truth convincing. Further, we do not see a motive to become convinced. We still think that the core meaning of truth in empirical matters consists of correspondence, but a correspondence that is divorced from metaphysical realism and open to many different universes of discourse. Of course, we need, in addition to an explication of the meanings of truth, also criteria of, indicators of, symptoms of, signs of, proxies of, and means to improve truth. Truth-by-operational-coherence may offer a very welcome additional criterion for truth that is applicable to some cases, but not a new meaning of truth. Chang, however, leaves no doubt that this is not what he intends and that it is not enough (200–203).

A key part of Chang’s project is to present us with a new scientific realism, one that is fitting for a pragmatist, and the sort of person who has worked in the philosophy of scientific practice. Chang describes in some detail the features of his view. On the one hand, he claims that “reality does not obey our commands” (71). This sounds like a firm commitment to realism, though many anti-realists would certainly endorse this claim as well. For example, nothing in Bas van Fraassen’s classic characterization of anti-realism implies that reality conforms to our theories (see van Fraassen 1980, §1.2). Even Thomas Kuhn was quite insistent about this. In his words, “you are not talking about anything worth calling science if you leave out the role of [nature]” (Kuhn 2000, 317).

But, then on the other hand, Chang claims that we must avoid the fallacy of pre-figuration (74). That is, we must not assume “that reality has well-defined parts and properties that exist independently of all conceptualization” (74).

This attack on the alleged existence of pre-figured objects of knowledge is central to Chang’s critical project (see 81). He insists that the widespread commitment among philosophers of science to the existence of pre-figured objects of knowledge is tied to wrongheaded notions of Reality and Truth, discussed above. These wrongheaded notions have prevented philosophers from making progress on key issues. And they have also led philosophers to engage in pointless debates with each other, debates that show little sign of ending. This attack on pre-figured objects of knowledge, though, sounds less like realism, and more like some form of anti-realism to us. Indeed, there is a Kuhnian ring to Chang’s attack on pre-figured objects of knowledge. This element of Kuhn’s view is most explicitly articulated in Paul Hoyningen-Huene’s exposition of Kuhn’s view (see Hoyningen-Huene 1993). To be clear, we are not objecting to Chang’s attack on pre-figured objects of knowledge, but rather noting a tension between his attack and scientific realism as generally understood.

We doubt that many contemporary realists, those actively engaged in the contemporary realism/anti-realism debates, the likes of Stathis Psillos, or even Steven French or John Worrall, would regard Chang’s position as a robust form of realism. Thus, it seems that the description of the position that Chang presents is not a form of realism that any committed contemporary realist would accept. French, for example, would think that the structure that our mathematical formulas capture does exist independently of all our conceptualization. And Worrall would insist that the reason Maxwell’s equations have persisted through changes of theory is because they are capturing something that exists independently of our conceptualization, specifically, the structure of reality. Psillos would go even further, much further. These realists are not going to give up on pre-figured objects of knowledge so readily, and they would certainly insist that a commitment to such things is integral to any position worth calling ‘realism’.^Footnote10

From their perspective, Chang has effectively abandoned the key elements of the most influential forms of realism accepted today. Granted, realists have softened their position over the last few decades, moving closer to views accepted by anti-realists. But the view that Chang presents is such a ‘lite’ form of realism that no contemporary realist would want to have it—it will not quench their thirst for real realism.

Indeed, it is useful to think of Chang as articulating a position on a continuum of positions from hardcore convergent realism to hardcore instrumentalism, according to which our theories make no pretentions to being true or are even properly evaluated as either true or false. The move towards the middle has been quite striking. In fact, many have noted how close Worrall’s structural realism is to van Fraassen’s constructive empiricism. Chang, it seems, is trying to articulate a position somewhere in this now somewhat crowded middle space.

But we do not think that Chang has to worry, or needs to make further concessions to our realist colleagues. The position that he has presented and described as scientific realism is a perfectly respectable position. In fact, it already has a name—anti-realism. Chang, we believe, is a scientific anti-realist. In fact, from our point of view, Chang even seems to have always been an anti-realist. Often, when we have heard his presentations at conferences, we felt his arguments supported anti-realism, whether he said so or not. And reflecting on the various dimensions of his philosophy of science, it looks thoroughly anti-realist to us. Indeed, many of his arguments, including his robust defense of pluralism, seem to fit nicely with anti-realism.

Let us offer some additional evidence for our claim that Chang is really an anti-realist. His enthusiasm for the pragmatists betrays an anti-realist strain. In particular, think of William James. James’ visit with Ernst Mach has been widely discussed. Apparently, they got along smashingly—two happy anti-realists talking about science for hours (see, for example, Frank (1949, Ch. 3). Both regarded scientific theories as instruments of sorts, the sort of thing we use rather than the sort of thing we know. This seems to capture an important dimension of Chang’s own view.

Thus, on our assessment, Chang was, and he continues to be, an anti-realist. Most fundamentally, he believes that scientists can accept theories that say contradictory things about reality and have no shame about it. According to Chang, we do not need to say that this is just a passing stage in the development of science, one that will be corrected in time, when we develop the one true proper theory. Nor do we have to be timid or secretive about it, as if working with contradictory theories indicates some sort of epistemic weakness.

In fact, reading Chang’s book, we were reminded of a quotation from a paper on positivism and realism by Moritz Schlick. In this paper, Schlick claims to be showing that “consistent empiricism does not deny the existence of the external world” (Schlick 1948 [1932–33], 505). That is, he is concerned to show that he and the Vienna Circle Positivists are realists. Like Chang, Schlick was really embarking on a public relations campaign of sorts, trying to reclaim the term ‘realism’, as it is a term with positive connotations that he wishes to associate with his own view. Schlick concludes his paper with the following remark (Schlick 1948 [1932–1933], 504–505):

The denial of the existence of the external world would be just as much a metaphysical statement as its affirmation. Hence, consistent empiricism does not deny the transcendent world, but shows that both its denial and affirmation are meaningless.

No one is really fooled by Schlick’s remarks and thinks that this demonstrates a commitment on Schlick’s part to realism as it was commonly understood in the 1930s. Rather, Schlick is affirming our convictions that Vienna Circle positivism is rightly characterized as a form of anti-realism.

Chang’s realism, we believe, is similar. The real realist, we think, is firmly committed to pre-figured objects of knowledge.

Underlying Chang’s arguments is a deep concern that we cannot afford to abandon the label ‘realism’. The battle is lost if we let the other side hold the ‘realism’ banner.^Footnote11 We can appreciate his concern, to some extent. There is an important public relations dimension to the rhetoric used in public debates. But we are not convinced that the most effective way to manage this is to insist on having one’s own view labeled ‘realism’. Rather, it might be a useful strategy to begin to teach the public that their expectations of science have to be more measured and modest. Chang is proceeding as if the right tactical move is to change the philosophers’ minds, to reform them. We think the more fruitful tactic may be to change the public’s mind. They need to realize that science, especially at the research frontier, and in practice, is a far messier business than is commonly believed.

We want to briefly return to Chang’s attack on Truth and Reality. Here he seems to be embarking on the same project that concerned Richard Rorty. Rorty was one of the most committed pragmatists in recent times. But his attempts to show his fellow philosophers that they were mistaken with respect to their views on Truth and Reality affected no change. Rorty is as close to a modern Socrates as we get. Though he was not made to drink the poisonous hemlock juice, he was pushed to the margins of the field. We would hate to see Chang meet a similar fate, and we doubt he will, given that he is always welcome amongst and appreciated by anti-realists. Some of the strongest lines of argument in Realism for Realistic People provide compelling grounds to accept anti-realism.

As we stated in the introduction, Chang defends mainly two positions in his book, pragmatism and pluralism. It is obvious that Chang really defends pragmatism in his book, he is not just considering or analysing it. Now, one may sense a certain tension between these two positions because obviously, pragmatism is one philosophical position among many others, thus his endorsement of it seems to amount to a monism. However, one may moderate or even eliminate this tension by realizing that Chang’s pluralism plays out on the object-level, regarding conceptual, epistemic, ontological, progress, and truth issues, whereas his monism concerns the meta-level. Thus, using the proven recipe of distinguishing the object-level from the meta-level, Chang may claim to have avoided logical contradiction. However, avoiding logical contradiction is not everything in philosophy, and at least to us, a certain uneasiness remains. If one looks at the arguments that Chang uses to defend his pluralism on the object-level, the question may present itself why these arguments should not equally have some force on the meta-level. Here are some quotes:

“Realism is generally taken as a monist position both in metaphysics and in the philosophy of science, but in my view monism must be rejected if it stands in the way of progress” (215). Replacing “realism” by “pragmatism” in the above quote, would also result in a very plausible statement. “True realism will pursue knowledge […] freed up from the unnecessary constraints of monism. All theories and all systems of practice that facilitate successful activities provide ways of learning, and they should all be maintained and developed actively” (215). We can clearly subscribe to this statement, but think that also pragmatism can be an unnecessary monistic constraint. “In science, all modes of inquiry should be put to work in this relentless drive to increase and improve knowledge” (215). Yes, indeed, and the same holds for philosophy.

With regard to science, Chang states: “In the process of epistemic iteration, we knowingly start inquiry on the basis of an imperfect starting point, and use the outcome of that inquiry in order to improve its own starting point” (239). The same procedure should also be applied in philosophy. We may start philosophical enquiry with an imperfect position like pragmatism, and in the course of the investigation refine and restrict pragmatism, and then enrich our means of analysis by other positions. “Iterative progress can also result in the improvement of methods and principles of inquiry. This applies to all manner of rules that are used in science” (244). Yes, indeed, and we believe that this also applies to the meta-level, philosophy.

“I also think that an explicit adoption of pluralism […] can enrich and ease the pragmatist view on truth. First, adopting pluralism concerning the concept of truth liberates the pragmatist theory of truth from having to account for all uses of ‘truth’ by means of one idea” (200). Well, it is not entirely clear how pluralist Chang really is when it comes to the execution of this idea. Remember that Chang stated the following: “But I do want to propose that truth-by-operational-coherence is what constitutes primary truth in empirical domains,” and these empirical domains include “science and much of daily life as well” (167). He really means all empirical domains, exceptionless: “I do want to offer a conception of truth that can be usefully applied in whichever domains that one may treat as empirical” (167). This means that in all empirical domains the other notions of truth can only produce secondary truths, and their “truth derives from the truth of other propositions” (165). Secondary truth, in other words, cannot stand on its own feet. It collapses if its referenced primary truth dissolves. This is not really pluralism but a master–slave relationship: the master eats whatever and whenever he pleases, but the slave must eat whatever the master hands over to him, otherwise he dies. So, in all empirical domains truth-by-operational-coherence is the only master truth game in town. How about the other domains? The only challenging domains that are left are “the a priori domains of logic and mathematics” (167). But also there, Chang wants pragmatism to have the upper hand. “It would have been the ultimate prize for a pragmatist to argue successfully that even logic was only pragmatically justified, rather than being a set of eternally valid ‘laws of thought’” (65). Although Chang implicitly refers in this quote to Dewey, we think that he really endorses this view, resulting in the claim that pragmatism is really a philosophy of science with all-encompassing scope.

As a result, Chang’s pluralism has some cracks on the object-level, by the dominance of a pragmatist theory of truth in all empirical matters, and does not seem to exist on the meta-level. When challenged on the latter point at a workshop on his book in October 2022 in Nancy, France, Chang denied that he is a metaphilosophical monist: he is also a metaphilosophical pluralist. His book is about pragmatism only because pragmatism has yet been underdeveloped in the philosophy of science. However, his pluralism does not imply the renunciation of judgment on philosophical positions, as an extreme form of relativism might. Instead, a pluralist can defend and advocate certain specific positions, while allowing that other positions should also be given the chance of being defended, intending productive outcomes arising from the interactions between different positions.

However, this defence of metaphilosophical pluralism provokes the question, who in today’s philosophy is not a pluralist in this very weak sense, at least in democratic states? Who would publicly deny that other positions should also be given the chance of being defended? The latter would imply that certain philosophical positions should be forbidden, which would be an absolutely intolerable intervention by whatever authority into academic freedom. Thus, serious metaphilosophical pluralism entails not restricting oneself to work in only one philosophical tradition and to be prepared that it is very unlikely that one philosophical size fits all. Feyerabend was a consistent pluralist, both on the object- and on the metalevel. He may have vehemently defended one philosophical position on one day but defended a completely contrarian position on the next. We are not defending pluralism here because our criticism of Chang in this respect is entirely internalist. He claims to be a pluralist but we identify strong monistic undercurrents.

Although this discussion note was critical in tendency, it should not be misunderstood as a depreciation of this book. Chang’s book discusses important topics, and it brings to renewed attention of philosophers of science that pragmatism possibly has something worthwhile to offer, also with respect to philosophy of science’s vision (see Brooks 2025). As philosophy is an enterprise that is apparently incurably controversial, the controversies about the value of the correspondence theory of truth, including the form divorced from metaphysical realism, about pragmatist theories of truth, about appropriate forms of realism, about monism and pluralism will continue. Chang’s book is a very welcome and fresh addition to these discussions.