The aim of this post is to introduce some central arguments and themes in contemporary analytic philosophy by connecting them to some arguments of Descartes, who is typically recognised as the father of modern philosophy.
I want to do this for a couple of reasons. One us just that the arguments are neat and interesting and deserving of widespread attention, and linking them to Descartes, who many may have encountered in one way or another, is just a neat pedagogical wrapping.
The second aim arises from this — if you’ve encountered Descartes, either formally in class or via wiki or youtube or some other resource, you might be left thinking that his project was both a failure and a misguided one at that. My aim is to show this isn’t so: in an era of replicability crises and journal hoaxes, Descartes’s aim to find a secure foundation for the sciences is extremely relevant, and although we can’t follow in the particulars of his footsteps, we should be attracted by its general motivations. Moreover, reflecting on what a contemporary Cartesian project should look like can help us to see how the concept and role of knowledge has changed in Western philosophy and society more generally since the early modern period.
A researcher in southern Germany, Erne Reaster, famous for doing work that cleared up a few theoretical issues concerning the behaviour of supercool liquids, recently published an article about how to solve the problems of replicability, hoaxing, and other issues afflicting the production of science today. Reaster, who has a PhD from MIT and was recently awarded a prestigious research grant from the Dutch government, claimed that his system, which makes us of certain recent work in game theory and Bayesian statistics, will finally enable the doubts cast on modern science practice to be dispelled.
Descartes: Smart Dude.
I made that up. Sorry. But I did so for a very good reason: it can be hard to appreciate the greats of philosophy. It’s tempting to take an almost condescending attitude towards them, because more or less as soon as we’re introduced to them we’re introduced to certain insuperable problems they face. Aristotle is introduced in one breath and said to be superseded by advances in science and logic in the next; we’re told of Kant’s critical project with the parenthesis that, of course, given developments in non-Euclidean geometry, it’s a non-starter. And soon after we learn about Descartes, we’re told about the infamous Cartesian circle which scuppers his project.
But Descartes was a smart cookie. And although it’s impossible to say, or even really make sense of, I think that were he around today he would be like our fictitious Reaster: a respected scientist whose goal was to reform the practice of a discipline in crisis. And, I hope you agree, condescension would not be the appropriate attitude to take to such a thinker.
Descartes, like Reaster, was a scientific prodigy, at the vanguard of discipline after discipline. Like Reaster, he lived in an era when scientific method was being impugned (albeit it instead of hoaxing and replication, he was worried that merely suggesting observation, as opposed to Aristotelian first principles, could lead to knowledge might have the Catholic church and their pitchforks knocking at your door). And like Reaster, he wanted to set scientific method on the right footing, using the best methods at his disposal.
But that was about 400 years ago. It was before calculus (indeed, Descartes’s work in pure maths was a precursor thereto), long before statistics, a time when Bayesian epistemology wasn’t a thing, indeed epistemology in general was barely a thing. And so Descartes started from a notably different position than our imagined Reaster.
In particular, he started, not with game theory and with statistics, but with himself and what he could and could not doubt. His aim was to find things of which he could be certain.
That’s reasonable enough, especially for someone with the mathematical background Descartes had, because mathematics deals in certainty. He thought if he could just establish a sound foundation of knowledge, then everything built upon it would be trustworthy. But what could such a foundation be? What could be known with certainty?
That’s not so easy to answer. Sitting in his room, he realized … well, he realized he thought he was sitting in his room, but sometimes people think things — because they’re deluded, or mentally ill, or perhaps dreaming — that are completely false. If he was dreaming, then he wasn’t sitting in his room. Since he couldn’t be certain he wasn’t dreaming (or hallucinating, etc.), he could doubt that he was sitting in his room, and since he was looking for something certain, he had to give up the thought that perception furnishes the foundations he was looking for. Fair enough. Maybe empirical knowledge is no good. But there are other sorts of knowledge: what about his mathematics, which doesn’t depend on experience in this way?
But, he says, it could be that an evil demon has tricked him, so that although it seems to him he’s certain that two and two is four, he’s actually deceived. So we can’t even believe in such necessary truths: again, even a smidgen of doubt, even owing to the outlandish possibility of an evil demon, is sufficient to impugn mathematical knowledge.
Despairing, he thinks to himself: I doubt everything! I know nothing. But wait a second. If it’s true that he doubts everything, then it’s true that he’s thinking, because doubting is a type of thinking. So if it’s true that he doubts everything, then at the time he’s doubting, it’s true that he’s thinking (even if he can doubt that he’s doubting, he’s still doubting, and so still thinking). And so there seems to be some thing that by its very nature he cannot doubt, namely that he is thinking. And this is the famous cogito argument: cogito, he says to himself, which is Latin for ‘I think’, and so sum — ‘I am’. He can’t conclude that his body exists: only that as a thinking thing he exists, and so he is drawn to conclude that his mind and his body are two different things, because they differ in their dubitability.
From that foundation, he attempts to rebuild the knowledge he’s cast in doubt. It’s generally agreed that he fails mightily in this (partly because of the Cartesian circle, mentioned above, but which I won’t get into), but we should agree at least that his project was and is a worthwhile one.
There are a couple of things I want to note: there’s doubt, the fear that you’re misled in a massive way — the worry of skepticism brought up by the possibility that you’re dreaming, or the plaything of an evil supernatural creature. There’s the importance of certainty — of knowledge, we might say. And there’s the difference between mind and body — and in particular in the thought that we have some super special guaranteed access to the mind that differentiates it from the body.
Skepticism — difference between mind and body — special access to the mind. Before going on, think to yourself: which of these views are most likely to have been discarded and which retained by contemporary thinkers? Do you think philosophers still take seriously the possibility of evil demons misleading us, or there being an important distinction between mind and body, or that the contents of the mind are in some sense easier to know than the nature of the body?
I hope your answers are: no, no, yes. Because those are exactly the wrong answers, as I will now show by running through three interesting and influential arguments from contemporary philosophy. I begin with the third question.
Timothy Williamson’s 2001 book Knowledge and Its Limits is arguably the most influential work yet published in the 21st century (if you’re extremely ambitious and want to dive in to contemporary philosophy, I might recommend getting it and slogging through.) And one of its central claims is that we don’t, contrary to what we might think, have a very special epistemic grasp on the contents of our minds.
The argument’s target is the following (I here and throughout dejargonify and simplify, sometimes omitting important details for the sake of expository ease while still retaining enough detail to let you see both form and content of the arguments.)
(Luminosity) Many mental states are such that, whenever we’re in them, we can know we’re in them.
Let’s think about this for a while. We have mental states: fear, belief, desire, pain, and so on. And there’s something special about our own mental states as against the mental states of others or states of the external world. They’re closer to us: we can know them more easily, just by introspecting. It’s easier for me to know I’m in pain than for me to know you are. This ease of knowledge might led one to think that if you’re in pain, you ipso facto* know* you’re in pain.
Williamson’s argument challenges this. It’s not always true that when you’re in pain, you can know you’re in pain — even if you’re paying close attention to yourself, you’re not on opiates, you don’t have pain asymbolia or other such things. You yourself right now could be in pain and not know it.
How? The crucial assumption underlying the argument is this:
(Reliability) Your belief in something only counts as knowledge if, in similar situations in which the thing was false, you wouldn’t also still believe that thing.
To see this, compare these two people: one of them thinks it’s always 9:20am, and says so every minute on the minute. Another person looks at the clock every minute and says what time the clock reads, and believes it’s that time. It’s 9:20am. Both say that it’s 9:20am, but the first doesn’t know because, if it were 9:21am, they would still say it’s 9:20am. The second does know because, if it were 9:21am, he wouldn’t still believe that it was 9:20am. The second is a reliable time teller; the first is not.
Now, imagine this scenario: you wake up one morning at 9am, and you’re cold, but you gradually warm up so that by noon you’re warm. In particular, imagine that for all minutes m, you’re less cold at m+1 than you were at m. But also, as is plausible, assume that for all minutes, you can’t tell the difference between m and m+1: they feel exactly the same to you, temperature-wise. On this basis, assume the following:
(Premise) If, at m, you know you feel cold, then at m+1 you feel cold.
This is the big and crucial premise. Williamson’s justification for it introduces details I’ve suppressed, so hopefully this is a way of putting it: m+1 feels exactly the same to you as m, by the gradualness hypothesis. In any two situations which feel exactly the same to you, you believe the same things, otherwise you’re not epistemically reliable. So, in m+1, you believe that you feel cold. But if in m+1 you don’t feel cold, then you believe falsely, and if you believe falsely in a similar case, then you don’t reliably know. So this premise follows from the reliability assumption.
(Think about a variation of the clocks case to see this: Imagine entering a room that just tells you the hour on some device. It reads 9am; you say and believe it’s 9am. It seems reasonable to say that you know it; but if after a few minutes, without the clock changing, you suddenly say and start to believe it’s 10am, it seems like there’s something dodgy going on, and you’re not reliable anymore.)
With that premise, we can make trouble. We can assume that at 9am you feel cold. Assuming Luminosity, we conclude that at 9am, you know you feel cold. From the Premise, we conclude that at 9:01am, you feel cold. From the Luminosity assumption, we again conclude that you know you feel cold, and from the Premise, we again conclude that at 9:02am you feel cold.
It’s clear we can keep going all the way to noon, and we’ll then have the result that at noon you are, and know yourself to be, cold. But that contradicts the description of the case: at noon, you are warm. So the reliability assumption, as manifested in the premise, and the luminosity assumption, jointly lead to a contradiction.
So if we like reliability, we must ditch luminosity. And that means that sometimes we can be in a mental state, but unable to know that we’re in that mental state. We do not have special epistemic access to our own mental states, something that would have surprised both Descartes and, I think, most of us still today.
Descartes appealed to an evil genius who could make it the case that everything he thought he experienced was in fact illusory. You might think that this somewhat out there possibility is a bit unserious. But if you do take it seriously — perhaps because you watched films like The Matrix, which are guided by a roughly similar premise — you might think, well, we’re screwed. If we’re so deeply illuded, there’s no hope for us, epistemically speaking.
An ingenious argument by Hilary Putnam, however, challenged this (it is found in the first chapter of his 1982 book Reason, Truth, and History). I should say before going on that while both the luminosity argument above and the dualism argument we’ll consider below are live concerns, the sorts of issues raised here are less pressing in contemporary philosophy. But they were so roughly a generation ago, and it’s worth knowing them, even if not for their intrinsic interest (which is great, I think), but for their historical importance.
Putnam introduces the science-fictional hypothesis that we are brains in vats. Some dastardly scientist, somehow, caused a bunch of brains to be generated just like ours from some special material that he kept in a vat. When the brain was fully formed, he put some preservative liquid in the vat and connected the brain up to wires. These wires lead to a computer, and that computer simulates experiences just like ours. You are to consider that you are such a brain-in-a-vat: that the computer or phone in front of you is illusory, as are the trees outside your window, your body, and everything else.
Now, should we be worried about that? Set aside what we may call existential worries — like whether being envatted precludes one from having an enjoyable and worthwhile (albeit simulated) life. We’re talking about truth and knowledge — should we be concerned that it might be true that we’re brains in vats, and thus that everything we know about the world is false?
Putnam thinks no. Roughly speaking, it can’t be that everything that we know about the world is false because the product of the vat’s computer program because, if we are indeed brains in vats, the world is just is the computer program. To explain this a bit more clearly, it’s necessary to take brief detour through some philosophy of language.
In the 20th and on into the 21st century there have been two main theories about how language connects to the world. On one, a word, say ‘gold’, stands for a particular bit of reality (the sum total of the world’s gold), because there is associated with the word a description that describes that sum total, like the bright metal more malleable than any other. Call this the descriptivist theory. On the other theory, ‘gold’ stands for the gold because there is some causal connection between people’s use of the word ‘gold’ and the stuff in the world. Roughly, when someone first encountered gold they said ‘let’s call that ‘gold’’ pointing to an example of gold. They talked about gold to their friends, whose use of ‘gold’ became casually connected, at one remove, to gold. They then in turn talked to their friends, and so on, such that my use of gold, via a bunch of intermediary people, can be traced back to the original baptising of gold, and that link makes it that ‘gold’ means the stuff that was baptised. The key point is the causal point: words stand for the things to which they are causally connected.
A consequence of this is that you can only speak and think about something if you’ve had some causal contact with it, either directly (if you’re the one who stood in front of gold and gave it its name) or indirectly (if you belong to a linguistic community that bears some connection to the person who originally gave it its name.)
But now think about our brains in vats. What have they had causal connection with? Well, it’s unclear. We might want to say that what they’re causally connected to is certain images the computer program produces, or maybe the underlying code responsible, or maybe nothing at all. One thing is clear: brains in vats have never been in causal contact with the vats in which their brains are envatted. They have never escaped from their vats to realize their situation. That means that they can’t think or talk about vats. Just as there are out in deep space craters shaped like my face (probably — space is big) that I can’t refer to because I’ve never been in causal contact with them, so brains in vats, having never been in contact with vats, can’t refer to them. That means that when a brain in a vat says ‘I’m a brain in a vat’, they don’t say something true. At best, they say that they are a brain in a vat in the computer simulation *(just as, at best, when they say ‘I’m reading something’, they say that they are reading something *in the computer simulation). But this means, if you’re a brain in a vat, you shouldn’t worry about being a brain in a vat — you can know, just because of how language works, that it’s false.
You may not like this argument: many people think it’s a big trick, and that it doesn’t really do much to assuage the fear that one is in a skeptical scenario. But I hope you agree it’s interesting and it does show how considerations from developments in philosophy — in particular about how language works — have shed new light on age old problems, and how such problems have developed.
Descartes thought that because he couldn’t doubt that he was thinking, when he was thinking, that enabled him also not to doubt that he was a thinking thing. On the other hand, because he could doubt that he had a body, he was able to doubt that he was a physical thing, and was thus led to posit dualism: that he was both mind and body, which are two different things.
You might think that today, because we know so much more about the brain and about biology in general, and because we tend to give less credence to religious-sounding concepts like souls or spirits, that such a view would find little favour. The mind, you might think, is just somehow the body, that it arises from it and its operations can be explained entirely in terms of how the brain works, which can in turn be explained entirely in terms of physics, eventually. We don’t need to posit minds in addition to bodies, which would be weird entities hard to square with scientific progress.
But contemporary philosophy of mind takes quite seriously the thought that the mind might be special and might require, for its explanation, going beyond the properties the physical sciences countenance. And one of the central arguments for doing so is somewhat — only somewhat — similar to Descartes’s. This is David Chalmer’s famous zombie argument, presented in his 1996 The Conscious Mind and in several works subsequently.
The argument tends to be formulated using a somewhat technical apparatus of what’s called two-dimensional semantics. I’m not sure I’m entirely on board with the necessity of using such apparatus for this purpose, and because it would take a long time to explain, I won’t bother.
Which is okay, because the argument can be presented quickly and clearly enough without it. A zombie is a creature that is an exact duplicate of a human being, but that lacks consciousness. Zombies — as far as we know — don’t exist in the real world, but they could exist. How do we know they could exist? Well, we can imagine they exist. We can picture a world in which there are creatures just like us — creatures going about their Sunday, eating their dinner, talking to their friends, going for walks by the sea, and so on, but for whom, unlike us, those experiences aren’t accompanied by conscious experience. They don’t taste the sharpness of the tomatoes, the cold of the sea air, the emotional warmth of talking to someone you like. But if we can imagine that so readily, then we should think it’s possible.
But that causes problems. To see this, let’s say a bit more about materialism. Materialism is the idea that, once we’ve got the physical properties, we get the other ones for free. Biological, chemical, economic, and so on properties all depend on the physical properties; they are nothing over and above those properties.
This talk of nothing over and aboveness might strike you as vague. It is, but that’s on me. There is a lot (like, a lot a lot) of philosophical work on this topic, but I can’t get into it here. But here is one thing we can say: if you like materialism, then you should like the following principle. We can imagine possible worlds other than ours: worlds in which kangaroos fly, for example, or there is a one hundred mile tall building.
Now imagine two possible worlds in which the same physical properties are distributed the same way: if in world one there’s a particle with a given mass at point p, so there is in world two. If there’s a field with a certain force at a location l in world one, a field with the same properties is in the same place in two, and so on for every fundamental physical entity and property.
If you like materialism, you should say that both these worlds instantiate all and only the same properties of any type: they instantiate the same biological properties, the same chemical properties, the same economic properties, and so on, and they do so precisely because they instantiate the physical properties. That’s just a fancy way of saying, really, that these other properties are nothing over and above the physical ones.
Now to the problem: we’ve assumed that it’s possible for zombies to exist. Translated into possible worlds talk, we’re saying that two worlds share all their physical (and thus biological, chemical, etc.) properties, but in one there is, and in the other there is not, conscious experiences. That is to say, in one there is, and in the other there is not, consciousness properties. But then that means that some properties are over and above physical ones, namely conscious ones, and so that gives us a picture of a world with two types of properties, physical and consciousness one, and that seems kind of like the Cartesian picture.
So those are our three arguments. To repeat, these have each been very influential, and have been or are taken very seriously. They give a good outline of part of the landscape of contemporary analytic philosophy.
You might not like them. If you don’t, I encourage you to think about why you don’t like them, think how you would clearly object to them, and so on. If nothing else, doing philosophy like that is a fun mental exercise.
I want to end with — I think — a slightly original point. Descartes sought scientific knowledge, and thus led him to seek for certainty in himself, for sure foundations. This has been enormously influential, and arguably our very concept of knowledge owes much to him, even if we don’t realize it.
Maybe we need to change our concept of knowledge today. It certainly seems like our culture is one in which the status of knowledge is uncertain. We have replicability crises, and hoaxes and other fears about peer review, not to mention fake news and culture wars outside of academia. But maybe Descartes offers us a hint as to how to proceed: knowledge is scientific knowledge, and that, today, is the knowledge produced by the institutions of scientific learning. That is, it’s something crucially involved with institutions, rather than something that a guy in a room by himself can figure out.
If that is so, then we need to ask hard questions about the roles that institutions play in generating or suppressing knowledge. We need to ask about how the publish-or-perish culture encourages sloppy or downright fradulent research, how the mechanisms for preventing it are imperfect, how certain voices are excluded while others are magnified (say, by institutional privilege, or citation networks), and other such questions. That is to say, a Cartesian project for the 21st century would have to realize the social element of how knowledge is created and disseminated. Our modern day Descartes, Erne Reacher, would, in the service of setting science on secure foundations, be studying bias and privilege and power, and those who think that studying such things is a waste of time — is a pointless exercise in grievance studies — are not, as they might like to think, the rightful heirs to the Cartesian project of furthering scientific knowledge but are in fact hindering it.