Scientism is the claim that science is the only source of knowledge. This claim has been the subject of intense controversy for years, and it has recently re-emerged in public debate following the publication of Steven Pinker’s Enlightenment Now: The Case for Reason, Science, Humanism, and Progress. Admittedly, Pinker does not make this claim himself, but those who do are (mis)using his work in support of their claims, and the renewed controversy over the term provides us with an opportunity to revisit its validity. Representatives of the humanities, in particular, have had their feathers ruffled by the notion that empirical observation and hypothesis testing have a monopoly on rational inquiry as tight as that enjoyed by Andrew Carnegie on the steel industry in the 19th century; liberal arts need not apply. Much of the criticism of scientism has focused on the aesthetic poverty of abandoning the contemplation of Shakespeare for the study of synapses in humanity’s quest for knowledge of the world and of ourselves. These criticisms have some merit, but a stronger case against scientism can be made using the philosophical equivalent of the arid reasoning that scientists admire.

Consider the following all-too-common scenario: you are a cash-strapped graduate student, and you decide to make a little money by participating in a psychological study.During the study you are presented with a computer program with two inputs, identified (with a tip of the hat to Nelson Goodman) as ‘grue’ and ‘bleen.’ The purpose of the study is unstated, as usual, so you simply focus on engaging with the program. As soon as you select an input, the computer prints out a number. As you continue clicking the inputs, you notice that every time you click ‘grue’ the computer prints ‘1’ and every time you click ‘bleen’ the computer prints ‘0.’ After a few minutes of doing this with no change in the pattern, you formulate a hypothesis that the computer is programmed to always return a ‘0’ for ‘bleen’ and a ‘1’ for ‘grue.’ In moving beyond making observations of the past and forming a testable claim about the future, you’ve made an assumption that the latter tends to resemble the former. We can call the assumption you’ve made the Past-Future Thesis (PFT):

The more consistently something has behaved a certain way in the past, the more likely it is that the thing will keep behaving the same way in the future.

What happens if you don’t assume the PFT, and instead assume that the future might be radically and unpredictably unlike the past? To make this clear, we’re now considering what happens if you don’t believe any past behavior makes similar future behavior more likely – that even a million trials with one result don’t make it any more likely that the next trial will have the same result.

On the Benefits of Philosophical Instruction

“Justification for such claims. Getting higher test scores after studying philosophy does not show that higher scores are the result of studying philosophy.”

QuillettePreston Stovall

In this case, all of your reasoning up to and including the testable claim about the program becomes invalid. Since you have no relevant data on what the computer does in response to each input, you cannot formulate a ‘law’ that the computer will always return ‘1’ for ‘grue’ and ‘0’ for ‘bleen.’ Your past observations are irrelevant because ex hypothesi they do not make it any more likely that the computer will return the same outputs as before. In fact, you are in the same boat, from an epistemological standpoint, as if you had never made any observation of the computer’s behavior at all.

And without law-like statements, Karl Popper’s celebrated criterion of falsifiability in science becomes unintelligible. Only statements that make claims about the future can be falsified by new data. Law-like statements (e.g., “Inputting ‘grue’ will always produce an output of ‘1’”) do make claims about the future, but mere descriptions of the past don’t. In the absence of a link connecting the two, there is no more contradiction in the future being radically different from the past than there is in Venus’s surface temperature being radically different from Pluto’s.

We can conclude, therefore, that the PFT is indispensable to science because it forms the basis of empirical reasoning. If you doubt this, try to think of even one claim about the natural behavior of anything – water, electrons, unladen swallows, the Moon, etc. – that does not depend on past observations of the thing in question. Doing so is, of course, impossible. Indeed, the claim that you can find things out about how things work just by thinking about them in the abstract, without observation, is the very sort of deductive reasoning that pioneers of the scientific method like Francis Bacon and John Stuart Mill ridiculed.

But this assumption that the future must resemble the past isn’t self-evident. It doesn’t force itself upon our belief in the same way that the impossibility of a square circle does. And because it isn’t self-evident, we ought to ask how we know it to be true. It is here that scientism comes to pieces. The PFT cannot itself be justified empirically, because empirical reasoning, as we saw above, assumes the PFT.

To see how this works, imagine how someone might try to defend the thesis empirically. One might say, “The PFT claims that past behavior makes similar future behavior more likely. And if we look at the past, we see that the future resembles the past all the time, so there’s an overwhelming probabilistic case for the PFT.” The problem, of course, is that in appealing to what’s happened in the past as a guide to what will happen in the future, the would-be defender is assuming the very thing in question – that the future will resemble the past. There is no non-circular way to bootstrap the PFT even as a probabilistic claim, much less an infallible metaphysical principle, from past experience.

We are left on the horns of the following dilemma: either the PFT can be justified on non-empirical grounds, or it cannot be justified at all. If we accept the first horn, then we are conceding that scientific observation is not the only source of knowledge, and thus that scientism is false. This is unacceptable to the defender of this philosophy, so what of the other option?

The second horn of the dilemma clearly won’t work either. On scientism’s account, science is the only source of knowledge. We’ve just seen above that the PFT cannot be known scientifically to be true. Ergo, given scientism, the PFT cannot be known to be true at all. You have no more reason, in that case, to believe in the PFT than your crackpot co-worker does to believe that Freemasons are controlling the world through a long-plotted 1960s takeover of the Vatican. In fact, you have less reason to believe in the PFT than your co-worker does in his Freemason Thesis, because he can at least point to a few conspiracy theory attempts at ‘evidence,’ but the whole enterprise of providing evidence at all has been shuttered for you by scientism’s restriction of knowledge to the empirical.

It gets worse, however. If you have no reason to believe that the PFT is true, then it follows that you have no reason to believe that anything that assumes the PFT is true either, and so out go all of your grounds for trust in any of the deliverances of science. Your belief that the Earth completes an orbit around the sun once every 365 days is no better grounded than the belief that the Earth completes an orbit around the moon once every 365 years – both are now equally unfounded and arbitrary.

It won’t do to say that we have no reason to believe the PFT to be false, and shouldn’t abandon it just because there’s a sliver of probability that it might not be true. This line of reasoning assumes that the onus is on someone who denies that the past is like the future to prove that view. But this is just creeping circular reasoning again, albeit better concealed: we find this argument tempting because it seems overwhelmingly probable that the PFT is true. And that sense of probability is, again, grounded in an implicit belief that it’s true because we’ve seen it be so in the past – and brings us back to the dilemma described above. If we actually adopt the sort of open-minded approach to this thesis that we would to any other claim, then the burden of proof falls upon the person who asserts, not the one who denies. Hence, this evasion likewise fails.