It’s by now a familiar argument: thinkers of what might be called a ‘relativist’, ‘postmodern’ or ‘anti-realist’ bent get accused of having corrupted science and society, leading to naked ideology in the universities and a blatant disregard for truth in the public sphere. While few will believe that Trump and his followers have deeply studied the works of Richard Rorty, Jacques Derrida or Bruno Latour, nevertheless some kind of influence is asserted, even if rather indirect (perhaps by weakening the ability of academia to fight back against post-truth phenomena).

There are really two questions here. One is whether these thinkers have caused a decline of truth. Another is whether their thought intrinsically leads to a decline of truth. The second question is about the content of their thinking, whereas the first is about what other people have done with their ideas — and that could of course involve misunderstanding, misrepresentation, misapplication. Answering the first question can only be done by detailed case studies: how exactly has Derrida been taken up in literary studies? How have historians of early modern Europe understood and applied the ideas of Latour? I will leave this question unanswered, mostly because I do not have the necessary expertise to answer it. Instead, I want to look at the second question. And I will do so at a relatively high level of generality. Instead of zooming in on particular thinkers, I want to discuss the general anti-realism which many of them advocate. I will argue that it is not opposed to truth at all, but that it can easily look to be so opposed when these works are read outside the philosophical context in which they were produced. (At the end, I briefly return to Rorty and Latour.)

I will introduce two characters, the Realist and the Anti-realist. Although their names suggest opposition, they are also allies against a common enemy, the sceptic. Where the sceptic believes that we cannot attain truth, both the Realist and the Anti-realist believe that scientific investigation reveals the world. Where they differ is in how they understand the logical relation between the two parts of this claim. For the Realist, the logically more fundamental notion is ‘the world’. We have an idea of the world, of reality, and then we understand scientific investigation as that activity that reveals this world to us. For the Anti-realist, the logically more fundamental notion is ‘scientific investigation’. We have an idea of how to attain the truth — that is, we have certain epistemic norms — and then we understand the world as that which is revealed by this truth-seeking activity.

Before talking about empirical science proper, I will illustrate this difference by applying it to the realm of mathematics. Take some unsolved problem in mathematics, for instance, the twin prime conjecture. Twin primes are prime numbers that differ by 2: for instance, 3 and 5, 5 and 7, 11 and 13, 17 and 19. The twin prime conjecture states that there are infinitely many such pairs. As I write this, this conjecture has been neither proven nor disproven. Now suppose that it cannot be proven and cannot be disproven. This is a possibility one must take seriously, since first of all, there is no guarantee that every mathematical statement can be either proven or disproven, and furthermore, it is known that there are statements which can be neither proven nor disproven (though this is not one of them). What then? Would the twin prime conjecture still be true or false?

Here is what the Realist would say: “The realm of mathematics is what it is independently of our procedures for proof. Either this realm contains infinitely many twin primes, or it does not. If it does, the twin prime conjecture is true. If it does not, the twin prime conjecture is false. It must be true or false. It must be true or false even if it can never be proven to be true or false.”

Here is what the Anti-realist would say: “In mathematics, to be true is not to be proven, because of course there are many truths that we have not yet proven. But to be true is to be provable. That is the only notion of mathematical truth that makes sense: whatever can be proven from the axioms is true. It is a mistake to believe that there is some mathematical realm that exists independently of our mathematical practice. Any statement that can (in principle) be proven is true; any statement that can be disproven is false; and any statement that can be neither proven nor disproven is neither true nor false.”

Clearly, what distinguishes the Realist from the Anti-realist in mathematics is not that one of them believes in truth and the other does not. Both believe in truth. They in fact believe in the exact same mathematical truths. But the Realist anchors truth in a realm of mathematical objects that can be understood apart from our ways of assessing mathematical truths, whereas the Anti-realist anchors truth in our ways of proving mathematical statements.

There are various reasons one might have to adopt an Anti-realist attitude. One might believe that understanding a statement is understanding how to find out whether that statement is true or false. Statements which cannot be shown to be true or false then turn out to be literally meaningless. (This is close to what Dummett might say.) Or one might believe that our idea of reality is inextricably bound up with our epistemic norms, such that the phrase ‘the realm of mathematics’ cannot be understood otherwise than ‘that which our mathematical investigations reveal’. Realism might then be chastised for positing some kind of inaccessible Platonic heaven of mathematical objects.

Let us now move from mathematics to empirical science. Could there be truths about the world that are — not just in practice, but in principle — beyond our powers of investigation? For instance, could there be a type of particle that exists, that is real, but that doesn’t interact with any other particles and therefore can never be measured? Here the Realist will say yes, while the Anti-realist will say no. For the Realist, we can first posit reality and only then ask the question of how to investigate it. For the Anti-realist, to be real is to be what we would encounter in a perfected investigation of the world. (For most Anti-realists, to be real is not to be already encountered, nor to be encountered by using exactly the strategies and instruments of investigation that are already in use. It is to be encountered in a perfected investigation; in the best investigation we are in principle capable of.)

The Realist and the Anti-realist do not differ in the importance they attach to the notion of truth. Nor do they differ in which truths they accept. They differ in their philosophical analysis of the nature of truth and the nature of reality. The Realist thinks that our concept of reality is logically prior to our concept of an epistemic norm, while the Anti-realist thinks that our concept of an epistemic norm is logically prior to our concept of reality.

(Caveat: Realism and Anti-realism have many meanings, and there are meanings where the Realist and the Anti-realist do differ in terms of which truths they accept. A Realist about superstrings believes that there really are superstrings, the corresponding Anti-realist believes that superstrings are merely a theoretical tool for explaining certain observations. While interesting, I don’t think this distinction, which hinges on the difference between observable and unobservable entities, is very useful for understanding the decline-of-truth narratives that we started out with. The problem with Trump is certainly not that he objects to unobservables.)

In a philosophical discussion between the Realist and the Anti-realist, the Realist might say: “Truth is correspondence with reality.” The Anti-realist might instead say: “Truth is what one arrives at when following our epistemic norms.” Now if one is not well-versed in the philosophical context, it may sound as if the Anti-realist is denying that truth is correspondence with reality. And then it may seem to follow that the Anti-realist believes that truths do not have to correspond with reality; that truths are either independent of reality or can even be the opposite of reality. And then it is easy to see the Anti-realist as opening the door to post-truth thinking. But the Anti-realist does not deny correspondence. It’s just that the Anti-realist believes that truth-as-correspondence, while of course correct, is not the fundamental analysis of truth.

Now one type of Anti-realist, no doubt the least threatening type, believes that there is one universal set of epistemic norms and therefore one universal idea of reality. Anti-realism perhaps takes on a more threatening look when one notices that epistemic norms change over time and that the norms embraced by various subsets of scientists differ. Is this then the idea that turns Anti-realism into a nefarious underminer of the love of truth?

There is a lot I could say about this (for instance about conceptual schemes and incommensurability), but I will restrict myself to a relatively simple line of argument. It is good to start from two things that everyone can agree on. First, different social groups hold and have held different epistemic norms. This is a fact that even a cursory glance at the history of science teaches us. Second, every group has always believed that the epistemic norms they held were binding on them; that is, they believed that these were the norms they had to live up to. This is also an obvious truth, because holding a norm is (perhaps among other things) believing that the norm is binding on you.

Where disagreement starts is when we consider what makes the norms binding on us. Here the Realist, for whom reality is logically prior to the norms, will usually say that those norms are binding on us that (most perfectly) reveal reality to us when followed. The Realist will thus posit a non-human authority, reality, that has authority over us in the sense that we have to consider ourselves bound by norms that are grounded in reality. The Anti-realist, on the other hand, for whom the norms are logically prior to reality, will usually say that we ourselves are the ground for the authority of the norms. This is not to say that the norms can be changed at will: they might be the result of rational self-insight, systematic experimentation, and/or a long historical development, and they might be highly resistant to change. But the norms cannot be justified by appeal to a non-human authority.

Both positions generate philosophical puzzles. The Realist will have to explain how we can know the norms, given that the norms are grounded in reality but reality can only be accessed by following the norms. There’s an obvious circle here. In addition, there are deep problems about norms grounded in an presumably non-normative reality. On the other hand, the Anti-realist must explain how normativity can be grounded in ourselves without degenerating into arbitrariness. There’s an enormous amount of philosophical territory to explore here, and several vast bodies of literature, but luckily, we can ignore it for now. I take it that the worries with which we started our investigation are not about technical problems within philosophy.

We pose again the question: does Anti-realism lead to a disregard for truth? I take it that the worry is not so much about being wrong — serious scientists have been wrong about many things, but they did not disregard truth — but about abandoning the very idea that there are epistemic norms we should live up to. We are worried about people who just say whatever they want, whatever is ideologically convenient for them, and don’t even ask whether what they say is true. Now Anti-realism is the claim that epistemic norms are logically prior to reality. That is not the claim that there are no epistemic norms. Indeed, if there were no epistemic norms, there would be no sense in having a debate between Realism and Anti-realism, since both are theories about the relation between epistemic norms and reality. If there are no epistemic norms, Realism and Anti-realism are both wrong.

Anti-realism can look like the claim that there are no epistemic norms only if one believes that norms must be grounded in reality. “If God does not exist, then everything is permitted.” But we have had little trouble accepting legal normativity without believing in the divine right of kings, or ethical normativity without having solved the Euthyphro dilemma. Even if one thinks, because of certain philosophical arguments, that Anti-realism is wrong, it seems rather extreme to believe that Anti-realism is so incoherent that it immediately leads to abandonment of the very notion of norms. Even a staunch Realist can accept that the Anti-realist in good faith believes that there are binding epistemic norms that are not grounded in a logically prior reality.

But — and I will now formulate what I think is a very common worry underlying many heated debates — if one believes that there have been many sets of epistemic norms throughout history, and even now there are many such sets; and one believes that these norms are not grounded in an extra-human authority; how can one still accept one’s own norms as binding? Must one not see all norms as equally binding? And since these various norms are inconsistent, must one not deny the bindingness of all norms?

Well, must one? What is this ‘must’? This ‘must’ appeals to some norm which is supposed to be binding. What norm could that be? Clearly, only a norm that we consider binding on ourselves, that is, one of our norms. Now, granted, if our norms imply that all norms ever held are equally binding, then we are in trouble. But we’re in trouble because we have adopted terrible and inconsistent norms. And of course our norms imply no such thing. Our epistemic norms do not imply that the epistemic norms of 17th century natural science were just as good as the norms we currently have. In fact, our norms imply that our norms are better than those of the 17th century. This is obvious.

The Realist and the Anti-realist both accept our current scientific norms. In that respect, there is no difference between them. Now it could be thought that the Realist, unlike the Anti-realist, has a story why our norms are better than those of the past: because they bring us closer to reality. But of course we evaluate that claim by applying our own norms — we take our norms, and we take what our norms reveal reality to be, and we conclude that our norms are closer to reality than the norms of the past. This is not only a rather uninformative procedure, but it is also one that the Anti-realist can perform in the exact same way that the Realist can. Of course the Anti-realist will have to use our current norms to do so, but our current norms are the norms of the Anti-realist. All our judgements are what Rorty calls ‘ethnocentric’: based in our current historical situation. But that doesn’t make them any less our norms.

(There’s a memorable passage somewhere where Rorty responds to the attack that he can’t say that the Nazi’s were bad. His response is that sure, the nazi’s were very bad. According to whose norms? Well, his norms. What other norms should he be using? Those of the nazi’s? This reply is entirely cogent and disappointing only to someone who was hoping for a big metaphysical proof that our current standards of morality are the correct ones.)

Furthermore, the Realist and the Anti-realist both accept that our current scientific norms are imperfect. It is easy to overlook, but normative systems are not static. They include norms for the improvement of norms. Finding out about the placebo effect led to the norm of using blind trials in medicine. This is an improvement even from the perspective of the earlier norms. For this reason, Rorty (at least in his more careful articles) talks about truth being what is revealed by the epistemic norms of beings whom we can recognise as better versions of ourselves.

We are now coming close to a conclusion. Realism and Anti-realism are very interesting positions from the perspective of philosophers, since they are fundamentally different ways of thinking about truth, reality, norms and the relations between them, with big consequences elsewhere in philosophy. It is much less obvious that they are relevant to the practical pursuit of science. The Realist and the Anti-realist accept the same epistemic norms, namely, the norms of current science. They agree about the importance of truth. They agree about the need to keep improving our norms. One may perhaps expect the Anti-realist to be more amenable to conceptual engineering and large-scale experimentation with norms, and the Realist to be more conservative, but that would be more a psychological phenomenon than anything that follows directly from the philosophical positions themselves.

(Hypothesis: the reason that Anti-realist philosophers have been misunderstood so often is that people thought their claims must have been directly relevant for science. And then it can perhaps start to look as if they say that you shouldn’t pursue truth, because what else can criticism of correspondence mean? But in fact a scientist could read Rorty and be convinced by Rorty and continue doing science in the exact way they always did, and that would be precisely what Rorty would expect. He is not advocating a change of scientific practice.)

Let’s consider two worries that might remain.

“I have heard Rorty say that truth exists only within a vocabulary. Surely this means that everyone has their own truth, in other words, that everything goes.” It is important to notice, first, that when Rorty speaks about a ‘vocabulary’ he is talking about a set of epistemic norms. Once you adopt, or rather, have a set of epistemic norms, there are clear ways of distinguishing the true from the false. I say ‘have’ rather than ‘adopt’ because Rorty doesn’t think this is a matter of choice. Norms, like languages, are the results of long historical processes. Now what Rorty denies is that there is a way to take two competing sets of norms and decide in a neutral fashion which is better; he denies that there are universal meta-norms. One may disagree with this for philosophical reasons, but it is not a very threatening conclusion. No one believes that there are unanswerable arguments that will succeed in converting a young earth creationist to science in practice; so Rorty’s claim that there are also theoretical reasons for believing that such arguments do not exist doesn’t make much of a practical difference. (Recall the passage about the Nazi’s above. Rorty doesn’t believe there is an ethical argument against Nazism that the Nazi’s are logically forced to accept. But of course nobody believes that there is an ethical argument against Nazism that they will in fact accept, so again it doesn’t make much of a difference.) I also want to add that Rorty thinks that very often, we can recognise certain norms as better than the one’s we currently have; he doesn’t claim we are stuck in the present.

“I have heard Latour say that an ancient Egyptian couldn’t have died from tuberculosis, because tuberculosis had not been discovered back then. Surely this makes a mockery of science, as if we invent rather than discover the world.” I don’t consider myself a Latour expert at all, but here’s my take on that. As far as I understand Latour, he is impressed by the fact that epistemic norms are much more specific and local than we usually believe. We like to use big, abstract words like ‘evidence’, but scientific practice is messy, complicated and often ad hoc. It therefore makes no sense to speak about bacteria in abstraction from a quite detailed story about how we detect them in scientific experiments. And so, yes, of course we can say that an Ancient Egyptian died from bacteria, but only once we have given a quite detailed story about how we could provide scientific evidence for this claim. (If we don’t know what would count as evidence for the claim, the claim is empty.) Since scientific practice could also have developed in very different ways, we should resist the temptation to think of bacteria as having always been out there, waiting to be discovered. There may be one sense in which that is true — we can do an experiment in which we find bacteria that were waiting in the mummy for ages and are now detected — but there is another sense, a scientifically less useful but philosophically more pregnant sense, in which it is false. It is not the case the science was bound to arrive at the theory of bacteria and these precise ways of detecting them. We could have ended up with a very different theory and very different experimental methods that would still have made sense of the world around us.

I add again that I’m not a Latour expert, so this might be mistaken. But it fits some things I’ve read of his, and it makes sense as a particular kind of Anti-realism, namely an Anti-realism that believes epistemic norms to be highly local, laboratory-bound, and contingent. Whether norms are that local and that contingent is of course debatable, but it’s not a crazy idea.