1. What is empiricism?
Your first encounter with the term ‘empiricism’ in a philosophical context may well have been in Early Modern, where empiricism was used as a wide-net descriptor for a host of rather diverse thinkers – in my version of the class it was Berkeley, Locke, and Hume – who believed the foundation of knowledge to be experience, accrued via acquaintance with empirical things, chief among them the objects of sensory experience. It is contrasted with ‘rationalism,’ which sees the mind and its faculties of rational inquiry as the foundation of all knowledge claims.
When we talk about empiricism in this class, the above isn’t what we mean, exactly, but it’s not totally unrelated. Van Fraassen says the rationalism/empiricism distinction was primarily made as an easy way to taxonomize pre-Kantian thinkers, which then allows professors to present Kant straightforwardly as the synthesis of these two schools. That’s fine, but it doesn’t exhaust the word. There are plenty of major post-Kantian philosophers who are also labelled as ‘empiricists,’ despite not referring to themselves as such (e.g. Mill). Are they just different terms? Van Fraassen says no — when we call a view ‘empiricist,’ we just mean that it denies some things and explanations beyond experience. Basically, if there are some philosophical theories and views that are totally focused on trying to explain what things are and why they are the way they are, and which argue for the existence of things we cannot access, then we should roughly sketch empiricism as the opposite: any robust philosophical view or theory which rejects demands for ultimate explanation and rejects attempts to postulate things beyond experience.
Historically, then, empiricism identifies a trend more than any one view; wherever there is some strong metaphysical theory, be it Platonism, Cartesianism, or some sort of continental gobbledygook about Dasein, there inevitably rises some brand or another of empiricism, whose proponents dare to say ‘enough!’ and cast off the yoke of systematized metaphysics and putative objects, bringing philosophy, as Locke said, back out of the ocean of mystery and into the realm of human understanding.
What does this have to do with science?
As discussed in Chakravartty, scientific theories seem to posit a lot of entities, forces, and other things that we either can’t apprehend without the aid of instruments (atoms), or can’t apprehend period (causation). The realist view of science will say that a good number of these things, even though we don’t have direct acquaintance with them, are real, in the same way that you are real.
But if we understand science this way, aren’t we just doing metaphysics of another sort? After all, if I demand to know why I fall down every time I jump, it seems like I’m not just asking how I should think about why I fall – it also seems like I want some more fundamental, behind-the-scenes why, and if there’s something we theorize as causing me to fall, it’s attractive to say that this something isn’t just a useful fiction, but it’s also real; after all, it seems unlikely that theories of gravitation could make incredibly accurate and consistent predictions unless they’re tracking something that’s really out there, beyond us. It’s intuitively attractive, then, to say that the putative objects put forward by our best theories are real, but in doing this, we must understand the sense in which the structures and objects posited are metaphysical ones – they go above and beyond what experience alone seems to present to us. Realists, of course, think this is a feature, not a bug, and are happy to commit to believing in plenty of these putative entities.
The empiricist response would look like this: we might have a theory of gravitation, and it’s a good theory. But it’s only good because it describes the things we want it to describe, and does so pretty consistently and accurately. Even though it’s good, that doesn’t mean it actually answers the real, outside-our-mind why of what’s going on, and there’s no good reason for us to say that the things it’s talking about are really there – they’re useful inventions which capture all the relevant features of what’s going on, but we have no real way of checking to see if they’re right – after all, there are lots of other things we could make up that presumably explain why I fall when I jump pretty accurately – that doesn’t mean we should say they’re actually there, right?
Empiricism, then, describes a kind of scientific anti-realism which still talks about atoms, and quantum particles, and black holes, but whose advocates, when questioned more deeply, will claim to use these terms more for reasons of parsimony. It would be wrong, according to them, to say science can give us the tools to find things beyond experience – all it can give us are the tools to construct explanations which make sense to the human mind and adequately predict and track the relevant phenomena.
2. What does it mean to hold empiricism as a view?
So now we might say: “Well, I’m convinced. I’m an empiricist now, which means I believe the following: the foundation of all knowledge is experience!” It’s tempting to think this way: when we say we subscribe to a view, we often claim that taking the view just means saying ‘I believe X,’ where X is some factual statement.
Van Fraassen thinks this doesn’t work. Why? First, he claims the empiricist must make the above assertion without justification, since the justification for such a claim would either be a) that the only claims accepted by the empiricist are those grounded in experience (which is circular) or b) that there is something else grounding this justification (what grounds that something else? This is an infinite regress). Granting this, the problem with empiricism is as follows: The empiricist’s central belief is that we cannot have a priori proof or denial of a factual proposition about the world, but this central belief is itself such a factual proposition about the world. Empiricism, formulated so, is vulnerable to a reductio ad absurdum – it must admit the admissibility of contrary views (for example, how would an empiricist deny the claim that “some knowledge is grounded by things distinct from experience,” if they are only looking to experience to verify such claims?) while also holding the inadmissibility of such views as its central dictum. The ultimate conclusion is that empiricism’s central claim seems merely putative, with no critical bite and no way of placing itself above any sort of metaphysical theory, despite those views being the central object of its theoretical ire!
To escape this hellish pit we’ve dug for ourselves, van Fraassen thinks we have to redefine what it means to hold a position. What if being an empiricist wasn’t just synonymous with ‘believing in a bunch of propositions.’? Van Fraassen says that we should instead treat empiricism attitudinally; being an empiricist means that you have a certain attitude towards science and value it for certain ends, and it is this stance which determines which beliefs about the world you may presuppose, and which new ones you’re willing to accept. This looks like familiar territory, overlapping with Chakravartty. But why does this solve the problems from the last section?
Simply put, thinking of empiricism of a stance now means that its principal disagreements with metaphysics will not be factual disagreements, but rather disagreements based on non-cognitive things, like values. Now, there is room for meaningful disagreement. Van Fraassen cites Feyerabend in pointing out that this view of stances dovetails nicely with how we tend to conceive of the limits of scientific admissibility; the problem with flat-earthers is not that their claims are intrinsically inadmissible (after all, the empiricist may well say we can’t really know what shape the Earth is), the problem is actually that the flat earther has certain attitudes and values regarding how science is done, and there is indeed a possibility that these values might, if we disagree, put their views outside the realm of legitimate scientific inquiry.
3. What about non-empirical views?
The above analysis is somewhat specific to empiricism, since the adoption of stances was a move primarily made to evade a reductio ad absurdum argument particular to the dogmas of empiricism. Is there any reason to think, say, a devoted materialist should care to think of their views as reflecting a stance?
The materialist might say no, that materialism is just synonymous with the belief that ‘matter is all there is,’ or something to that effect. But van Fraassen thinks this is a form of ‘false consciousness’ – the materialist, even if they aren’t wrong per se, is misunderstanding what their own position is. The weird thing about materialism, as van Fraassen points out, is that, throughout the history of science and philosophy, there are always materialists, and yet materialists of yore often not only held different beliefs than the materialists that followed, but likely would have held as inconceivable the beliefs of their successors – would an ether theorist have simply balked at the possibility of their being an object such as a photon, or indeed, spacetime? Would a Newtonian-era physics have allowed for an account of quantum objects as material? It seems likely that, in it at least some cases, the answer is no.
So, does materialism, as a term, just extend over a concatenation of a bunch of discrete and incommensurable beliefs? Maybe. But we have some clues to indicate that this isn’t what’s going on. Look, for example, at how materialists react during scientific revolutions; when new scientific paradigms are suggested, the old scientists often embrace the new theories and continue to call themselves materialist, even though they have adopted a new set of views about material things whose putative properties might have disqualified them as ‘physical objects’ in previous formulations. Yet the materialist scientist doesn’t seem to see any problems here, and will likely offer one of two solutions provisionally proffered by van Fraassen.
The first option to say is “Well, matter is still everything, it’s just that my definition of ‘everything’ was expanded, and so it turns out that matter is a little different than what I thought.” The second option is to say “well, there’s no contradiction, because materialism is really just a completed physics. These new entities are a part of physics, and so, while they do stuff I didn’t previously consider options for material things, I see no problem with calling these new things material, in virtue of their being part of a (now more complete) physics.”
I hope it’s clear what’s odd about both of these responses; they’re putting the cart before the horse. The materialist position almost always seems to justify its claims not merely presently, but futurely. They are saying :“Not only are all putative objects posited by scientific theories physical ones, but future scientific theories will also posit only material objects, never immaterial ones.” Materialism is being presumed, not so much as a set of dogmas, but as a mold in which all new scientific developments and theoretical objects are taken and cast in some material capacity. Materialism, then, seems less like a set of defensible propositions, and more like an attitude or policy which determines not only how one treats their knowledge, but how one intends to treat future knowledge. The reasons for attributing physical-ness to ether, and light, and black holes, and dark matter, are at their core non-cognitive in origin – a stance.
Van Fraassen doesn’t give an exhaustive treatment of all possible positions, but empiricism and materialism, broadly construed, seem to capture the vast majority of debates going on in philosophy of science regarding what science can tell us (and regardless, his treatment of materialism applies just as well to physicalism and other similar-to-materialism-but-not-materialism positions). The implication that most or all philosophical positions on scientific inquiry are actually stances should reshape how we understand scientific disagreement and set the parameters of admissibility in the sciences.