1. The Gist of the Cogito
René Descartes is responsible for what might be the shortest and best-known argument (if we may call it an argument) in the history of Western thought: I think, therefore I am. This is often referred to as the Cogito, due its Latin form: Cogito, ergo sum.
I often run into misunderstandings of the Cogito, which Descartes first stated in his Discourse on Method (1637) and elaborated on in later writings. My aim here is to briefly explain what it means and, more importantly, what it does not mean—or, more specifically, what cannot be logically inferred from it. I do this over two sections, where the first section (i.e., this one) addresses what the Cogito means, and the second section treats the Cogito from a perspective of basic logic.
The Cogito is generally understood, roughly, as follows. No matter how much I may doubt about the world, one thing I cannot doubt—and indeed cannot be wrong about—is that I exist. Why? Because if thinking is happening, there must be some thinking thing doing that thinking. Indeed, some translations have it as “I am thinking, therefore I exist.”
“To think” here means to experience or sense or feel in some way that involves mental content, conscious awareness, and so on. And so the very presence of perceptual experience is enough to count as “thinking.” Which is to say that one need not actually have the thought, “I think” in order for thinking to function as a marker of existence. My favorite way to put the Cogito is: “I experience, therefore I exist.” Almost as good is: “I’m conscious, therefore I exist.”
Let me put all this another way. You might turn out to be a brain in a vat or stuck in the Matrix. The world around you and even your own body might be illusory (maybe you are some creature with 12 heads but are vividly dreaming that you have one head). You can, conceivably or in theory, be fooled about all that. But there must be some mind—or, more precisely: some experiencing entity, some entity with a mind or at least mental faculties—that is being fooled. “I am being fooled, therefore I exist” captures this point as a special case of the Cogito (in which the experience in question has the flavor of “being fooled”).
If such ruminations lead you to doubt your existence, you might conclude, “I am doubting my existence, therefore I, the doubter, exist.”
That’s the gist. A simple idea, but one that has been misunderstood by some very smart folks. Thomas Hobbes, for example, mocked it by responding that you might as well say, “I walk, therefore I am an act of walking.” Descartes’ response to this objection is well worth reading. You can do so in any edition of his Meditations on First Philosophy (first published in 1641) that includes objections and replies; I like this book (see page 76, Objection II): Modern Philosophy: An Anthology of Primary Sources, 2nd Edition.
This in mind, consider the following critical invocation of the Cogito in Eckhart Tolle’s popular book, The Power of Now (1999), in a chapter called “You Are Not Your Mind” (page 15):
The philosopher Descartes believed that he had found the most fundamental truth when he made his famous statement: “I think, therefore I am.” He had, in fact, given expression to the most basic error: to equate thinking with Being and identity with with thinking. The compulsive thinker, which means almost everyone, lives in a state of apparent separateness, in an insanely complex world of continues problems and conflict, a world that reflects the ever-increasing fragmentation of the mind. Enlightenment is a state of wholeness, of being “at one” and therefore at peace. At one with life in its manifested aspect, the world, as well as with your deepest self and life unmanifested—at one with Being. Enlightenment is not only the end of suffering and of continuous conflict within and without, but also the end of the dreadful enslavement to incessant thinking. What an incredible liberation this is!
Unsurprisingly, I’m unconvinced by Tolle’s critique, which has some echoes of Hobbes’ critique; here, though, Descartes is incorrectly accused of equating thinking with “Being and identity.” Rather, the Cogito claims that thinking provides indisputable evidence of the existence of some thinking thing (i.e., the thing doing the thinking).
Admittedly, I did not finish Tolle’s book. Maybe his critique of the Cogito is helpful for the sorts of problems his readers are trying to solve, despite failing to successfully demonstrate that thinking is not proof of an existing thinker—however fragmented or scattered or incessant that thinker’s mental goings-on and content may be, and even should that thinker, that experiencer, have no concept of “I” or existence.
Such points, some of which I may be unfairly ascribing to Tolle, do not bear on the argument or intuition (or however you’d like to classify the Cogito) that experience is a marker for the existence of some experiencing entity. For starters, whatever sorts of claims such observations allow, those claims would not amount to the counterexample needed in order to clearly dismantle the Cogito, which is, broadly, “here is a thinking thing that does not exist.” More on this below, when I consider the Cogito as a basic logical sentence. For now, I’ll reiterate that, strictly as a matter of interpretation, the Cogito does not imply an equality relation between existence—or as Tolle puts it “Being and identity”—and thinking.
That said, I suppose it’s safe to assume that Tolle does not raise the Cogito in order to challenge it on formal or analytic philosophical grounds, but rather for pedagogical purposes, in which case the example fits snuggly under the common theme of making the Cogito mean whatever you want it to mean.
That this happens with the Cogito is understandable, given that “I think, therefore I am” doesn’t contain the entirely of Descartes’ account, so much as slogan-ize it as a soundbite (thus my preference for “I experience, therefore I exist,” which is closer to the point for contemporary readers). As we’ll see in a moment, this is why I think the second part of this writing will be an easier one to undertake—in which we treat the Cogito at face value as a basic conditional statement. In which case the Cogito is subject to some straightforward rules of logic, the violation of which is less forgivable than is the general misinterpretation of Descartes’ arguments that surround, or are summarized by, the Cogito.
To be clear, there are things to worry or wonder about with respect to those arguments, even if we stick close to the Cogito proper, setting aside Descartes’ even broader project, involving, for example, the existence and nature of God. We might, for example, wonder whether a mind (or, even worse, some vague collection of mental faculties) is enough to ground the claim that I exist if the thinking thing—particularly if the entity has a meta-cognitive concept of self and of “I”—turns out to be drastically different from what it believes itself to be; if, for example, it believes itself to be the human actor Keanu Reeves but turns out to actually be a 12-headed creature being fooled by the Matrix, or turns out to be a swarm of insects upon whose swarmy activities consciousness has momentarily supervened. The thing being fooled here, the “it” being fooled here is the existing thing, but arguably the meta-cognitive concept of self, of an “I,” that existing thing has assembled (or been fooled into assembling) may very well exist in its own right, and so on and so forth, unraveling into a tremendously difficult discussion.
These worries about personal identity—about whether “I” indexes anything or nothing in particular, or may index many things—might be especially problematic if we, as mind-body monists, reject Descartes’ notion of the mind (i.e., the soul) as existing independently of the physical body. For example, suppose it turns out I am in no way a biological human, but rather the emergent epiphenomenon of an insect swarm. My use of “I” could be said to index the subjectively represented person I take myself to be, the swarm of insects, or both, depending on our metaphysics about the mind-body relation and personal identity. (This also brings to mind Hilary Putnam’s answer to Brain-in-a-Vat style arguments.)
I think the Cogito, in its basic function as a marker of some (at least one?) existing thing, survives challenges of the above sort (which I’ll briefly consider more closely in an Addendum [Section 4]), as well as other sorts of challenges, such as those pointing to a potential circularity or question-begging in the Cogito proper. You can explore well-known criticisms of the Cogito at Wikipedia and the Stanford Encyclopedia of Philosophy, offered up and debated by scholars who’ve thought a lot more about the Cogito than I have. But those scholars do generally agree on the gist of what the Cogito means.
That gist is what I’ve focused on here as the interpretive starting point for meaningfully challenging the Cogito. To be clear, my goal today is not to defend the Cogito, or even to persuade you that your interpretation must harmonize with the above-described gist (though I think it really must). My real goal is to show that even if you think the Cogito fails, there are fair and unfair ways to argue that it fails. More straightforwardly, there are logically valid and invalid ways to argue that it fails.
2. The Cogito as a Logical Sentence
Having given a small taste of what it’s like to engage the Cogito at its deeper interpretive levels, I’ll turn now to the easier job of outlining what can and cannot be logically inferred from it. I’ll cover three inferences one might be tempted to make from the basic statement of the Cogito. One of these is legal (the contrapositive), the other two are illegal (the inverse and the converse).
From what I’ve seen, it’s at this more superficial level of engagement that the Cogito is usually cited in public, non-academic forums.
2.1 The Material Conditional
To begin, let’s restate the Cogito in the form of a material conditional, or if-then, statement:
If I think, then I exist.
Formally, we can represent this as:
T → E
A few comments about the material conditional. In its generic form, it is typically represented as:
P → Q
I say this as “If P then Q,” but some people prefer to say is as “P implies Q” or “P entails Q” or some other way.
Technically, this is best viewed as the starting premise for an argument that next goes on to say: “I think, therefore I exist.” The form of this argument, taken all together, is known as modus ponens, but we needn’t go this far here. What matters here is what can and cannot be validly inferred from the the rule (or assumption) T → E, which is precisely the rule that grounds the Cogito. Discussion of this rule, then, gives us a tidy way of discussing the logical relationships at hand. I hope that’s clear.
To be clear, a rule stated in this way does not have to be true in order to validly ground an argument. Again, whether you think the Cogito true or false, the rules of logic hold. The study of logic is usually about validity, not truth (or what logicians call soundness).
It’s also important to notice that we can construct a large variety of statements with the material conditional connective (or operator). “If it rains, the ground gets wet.” “If Tonya goes to the party, then Oscar lost his bike.” “If the battery dies, the remote control won’t work.” “If the coin landed Heads, it didn’t land Tails.” “If I don’t die today, Henry won’t haunt you tomorrow.” “If the coin landed heads, the Earth didn’t explode.”
We needn’t worry about what makes the operator work (e.g., it needn’t have anything to do with causality). We’re only concerned here with what we can and cannot infer given a material conditional as our starting point. Everything here will be intuitive enough in that regard, but to drive it home, I’ll also use this clearer example:
If I’m a dog, then I’m a mammal.
D → M
2.2 The Contrapositive
Here’s the one thing I’ll review that we can validly infer. Its form is called the contrapositive (note that the symbol “~” represents the word “not”):
If I do not exist, then I do not think.
~E → ~T
The validity of this inference is not always immediately intuitive, but I think its validity is clear here. Still, let’s compare it to the even more intuitive analogy:
If I’m not a mammal, then I’m not a dog.
~M → ~D
I assume it to be obvious that if something is not a mammal, it cannot be a dog. But allow me to reemphasize that the dog-mammal analogy is only for bolstering intuitions. As I mentioned above, a conditional statement need not be true. “If I’m a mammal, then I’m made of cheese” would work just as well. And anyone who claims to believe that statement would be committed, at least implicitly, to the claim that anything that is not made of cheese is not a mammal.
This is, in fact, what I mean by “we can validly infer”—i.e., I mean that anyone who is committed to P → Q is also implicitly committed to the contrapositive of that statement: ~Q → ~P. At the very least, that person is not committed to any of the illegal variations of that statement. In other words, if you disagree with the Cogito, it is not a valid move to challenge it by way of an invalidly inferred reformulation, such as those I’m about to review.
I just did a quick look at some online forums at which folks are discussing the Cogito and easily found several instances of committing Descartes to such illegal variations. Of course, these mistakes arise often in all sorts of discussions that have nothing to do with the Cogito. Which makes them all the more worth talking about.
2.3 The Inverse
The first illegal move I’ll look at is the inverse.
In 2016, Bill Nye the Science Guy made headlines (in certain circles) by giving an undergraduate philosophy major some bad advice about the value of philosophy while showcasing an egregious ignorance of the field. Nye’s advice included this unfortunate evaluation of the Cogito:
I think, therefore I am. Well what if you don’t think about it? Do you not exist anymore? You probably still exist. Even if you’re not thinking about existence.
I’m not sure if Nye means that, according to Descartes, we must be thinking about existence in order to exist, or that we must be thinking period. As the Cogito says nothing about “thinking about existence,” I’ll charitably assume Nye means the latter (or at least a special case of the latter, since thinking about existence entails thinking). In other words, I’ll assume that Nye means to commit Descartes to the inverse of T → E, which is the following:
If I’m not thinking, then I don’t exist.
~T → ~E
In my experience, this is the most common fallacious interpretation of the Cogito. That it is unfair to commit Descartes to this proposition is especially clear given the analogy:
If I’m not a dog, then I’m not a mammal.
~D → ~M
Let’s explore a slight variation on the analogy. If someone claims to believe that if there’s a dog in the bathroom, then there’s a mammal in the bathroom, that person is clearly not committed to believing, upon learning that there is not a dog in the bathroom, that there is not a mammal in the bathroom. There could be a cat or a human in the bathroom.
For the same reason, Nye is wrong to commit Descartes to the belief that, if someone isn’t thinking, then that person doesn’t exist. The Cogito allows for the possibility that one may very well be unconscious while existing—while undergoing anesthesia, for example. But the Cogito also doesn’t imply that one must continue to exist while not thinking. Any objections you’d like to raise against the soundness of the statement “If I’m not thinking, then I don’t exist” do not apply to the Cogito, given that it is invalid to infer the inverse of a material conditional. Which is to say that ~T → ~E is not the Cogito, nor is it logically equivalent to the Cogito. (To mistakenly claim otherwise is often called denying the antecedent, or simply the fallacy of the inverse).
My impression is that Nye has done a lot of good for popularizing science. I hope he’s learning to do a better job of talking publicly about philosophy. See his full answer to the student, which has more wrong with it than I’ve addressed here, in this short (3:41 minutes) video on YouTube: Hey Bill Nye, ‘Does Science Have All the Answers or Should We Do Philosophy Too?’
2.4 The Converse
The next illegal move is to infer the converse:
If I exist, then I think.
E → T
Again, this obviously doesn’t follow from the rule T → E. Here’s the dog-mammal analogy:
If I’m a mammal, then I’m a dog.
M → D
This seems to be the most common mistake I encounter among mistreatments of material conditionals more broadly, particularly when they take a slightly different form, along the lines of “Since all or most A are B, all or most B must be A.” To see how wrong this is, imagine saying “Since all cantaloupes are fruit, all fruit must be cantaloupes.” One way to conditionally state this is, “If it’s a cantaloupe, then it’s a fruit.” From this, we obviously cannot infer—we cannot commit the believer in that statement to—the converse: “If it’s a fruit, then it’s a cantaloupe.” (This is often referred to as affirming the consequent, or just the fallacy of the converse).
This can be extended, by the way: “If it’s an A, then it’s a C; and if it’s a B, then it’s a C” does not imply that “If it’s a C, then it’s an A or B.” Nor, in fact, does it imply that “If it’s an A, then it’s a B.” It doesn’t even imply that some A are B. Here’s an intuitive example demonstrating this last point: “All insects are animals and all squirrels are animals, therefore some insects are squirrels.” (This falls under something often called the fallacy of the undistributed middle.)
It’s interesting to reflect not only the high frequency with which these basic fallacies come up in the socio-political discourse, but also on the fact that some folks deem it obligatory to point them out in some contexts while deeming it obligatory not to point them out in other contexts.
To summarize, Descartes’ Cogito essentially states that if you experience mental content—whether you are dreaming, are in the Matrix, are the emergent epiphenomenon of an insect swarm, are a Boltzmann brain, or really are just sitting at your desk right now eating a sandwich —you must exist. And if you disagree, for whatever reason, with that claim (often summarized as “I think, therefore I am”), or with whatever use you think Descartes meant to put that claim towards in his broader arguments—it would be invalid to pursue that criticism by way of challenging invalid reformulations of the Cogito, such as its inverse or converse.
4. Optional Addendum: Notes on Challenging the Cogito
I’d like to look a little more closely at the above-noted prospects of challenging the Cogito given certain metaphysical assumptions about the mind-body relation and personal identity. The aim here, again, is not to defend the Cogito. If anything, this closer look will pose a problem to arguments I’ve expressed elsewhere (more about which below). I won’t resolve anything today, but will raise some questions for potential future consideration.
(In fact, I’m throwing these on a last-minute. Unfortunately [or fortunately], I don’t have an editor to stop me.)
Before starting, let me emphasize that, in this section and in this writing overall, I’m talking about challenging the Cogito, and not about confirming, much less proving, it. What it takes to confirm a P → Q rule would involve a different and more difficult discussion than the present one.
The clearest way to challenge the Cogito would be to show an instance of thinking in which there is nothing that it is doing the thinking. That is, in which the antecedent T is true while the consequent E is false, as in: “I think and I don’t exist” or “here is a non-existent thinking thing.” To point to a clear, empirical instance of such a thing would provide a counterexample that disproves the Cogito. To clearly demonstrate its theoretical possibility would have a similar effect, or at least would severely challenge the Cogito.
Of course, what we’re ideally striving for is the empirical grounds of having found a counterexample. For example, the clearest challenge to the rule “If an animal is a swan, then that animal is white” is to point to a real-world creature about which you can correctly say, “That animal is a swan and is not white.”
What does not work is to point to a raven and say, “That animal is not a swan and is not white,” or to a dove and say, “That animal is not a swan and is white.” The latter two bits of evidence do not challenge the rule, “If an animal is a swan, then that animal is white.”
It’s important to note here that we must have some definition of what a swan is before we can produce a counterexample. It so happens that the bundle of characteristics we’ve determined merits the label “swan” does not entail the swan’s being white (in other words, it’d be overly arbitrary to include that characteristic for what I think are obvious reasons). It also just so happens that all the swans observed are white. Until we see one that’s not white. This is, of course, what happened in the real-world case in which the long-held assumption by Europeans that all swans are white was disproved by counterexamples in Australia in the 17th century.
How might we go about finding a counterexample to the Cogito? Here’s one way. But I should note that it’s only effective if you’re challenging the Cogito in the eyes of someone who claims to be a compositional nihilist (or holds some similar view, particularly with respect to brains and minds and selves and suchlike; read on to see what I mean).
I have, in fact, proclaimed myself to be a nihilist about composition (also known as a mereological nihilism). This essentially means that I don’t take composite objects to literally exist. For example, the chair I’m sitting in is made up of billions of smaller objects—of whatever the smallest possible things are, let’s call them particles—that are, to use the mereologist’s lingo, “arranged chair-wise”; but the chair itself does not actually exist as a singular thing in itself. In other words, I have a mental model of the chair and I use language and such to pick out general or particular chair-wise arranged particles (that are held together, for some duration of time by various forces, in a certain relation we call “chair”)—but no actual, literal, singular thing exists in that region of space occupied by any given set of chair-wise arranged particles.
It so happens that I extend this view not only to inanimate things like chairs, but also to organisms—including humans. There are other folks, however, who agree with me about inanimate objects, but not about organisms. The best known contemporary proponent of this “organicist” view is, as far as I know, Peter van Inwagen. His argument is essentially that because I do recognize myself to not be a simple object (which is to say that I do recognize myself to be a singular, composite object), full-on nihilism must be false.
I’ve argued against van Inwagen’s view here: “In Favor of Compositional Nihilism: A Response to the Organicism of van Inwagen.”
I’m not convinced that my argument there is successful (and I haven’t reread it since writing it), but my intuition that is successful persists. And I now find myself facing a further challenge involving the aforementioned metaphysical issues about the mind-body relation and personal identity.
Let’s assume (safely, I’m sure) that I really am, more or less, the biological human I take myself to be. How I parse out my “self”—that is, what sort of thing I am or what the “I” in “I think” denotes, etc.—is clearly not so easy to do, particularly once my mind is introduced into the mix. If I (as current dogma among many institutionally prominent scientists and philosophers etc. would have it) consider the mind and brain to be identical; and if I consider my brain to not really exist (rather, there are particles in my skull arranged brain-wise), while also considering my brain to be the center and source of my experience, then it just might be reasonable to say that I’m committed to the following view: there’s thinking happening, but no particular thing exists that is doing that thinking.
I’m inclined to endorse that view! But I also don’t think it poses a counterexample to the Cogito. Clearly I exist. And the fact that I have experience is, for me, the clearest empirical marker of that fact. My being able to reconcile all this (provided I’m not just imagining a tension where there isn’t one) will come down to how we define all the stuff involved in this picture (see again the comments on defining “swan”). Which is to say that I owe some explanation of what, then, the “I” in “I think, therefore I am” refers to.
I won’t attempt that here. But I will say that this difficult job might be made a little easier by the fact that, despite self-identifying as a physicalist, I happen to not take the mind and brain to be identical, nor do I take my self (which I do consider to exist, at the very least diachronically) to be identical to my (with heavy scare quotes around “my“) body; nor do I take my mind or self to be identical to any subset of my body’s physical parts. Exactly what sorts of entities mind and self and such are, I have not yet managed to understand, and likely never will. Nor am I certain of what status I’d give them should I come to believe that I am a Brain in a Vat or the emergent property of a swarm of insects. I’m not sure the problem in those cases is any harder than what I believe to be my actual situation, in which I am a biological human.
(Some folks might think the insect swarm example isn’t all that different from that of being the epiphenomenal, emergent property of a “swarm” of brain cells; see, for example, Alex Rosenberg’s 2011 book The Atheist’s Guide to Reality: Enjoying Life without Illusions for an account of what I take to be an epiphenomenal picture of consciousness, or is at least close to such a picture, though is of course much more solid than my tongue-in-cheek “swarm of insects” example. My problem with epiphenomenal views, insomuch as I understand them, is that they don’t seem to do enough to deal with mental causation, or what might be better characterized as a kind of feedback loop in which observation of one’s own mental content contributes to succeeding brain/mental states; this is yet another complicated topic for another day, but I will say that I find the input-output model described in Rosenberg’s book more satisfying than other accounts I’ve so far seen, if not sufficiently satisfying to convince me, namely due, again, to my not being sure how the account is meant to treat the input of one’s own experienced mental states.)
Though I can’t say exactly what sort of thing I am, I am sympathetic to arguments in which what I take myself to be now survives my coming to believe that the material world is far different than my current mental models have it to be, bringing into question the importance of the substrate underlying that reality (see again the above-noted arguments from Putnam). I don’t see this as being inconsistent with materialism or physicalism, though endeavoring to explain how would open several cans of worms I’d rather leave on the shelf today. But the shelf is right over there and well stocked and that day will come.
Now, van Inwagen’s approach here is to hold firmly onto the Cogito by classifying organisms as composites, unlike chairs which, strictly speaking, don’t exist (according to bother me and van Inwagen, anyway). While my approach has been to say that organisms (and, of course, brains) need not exist, strictly speaking, in order for it to be correctly said that there is some thing that is thinking—for there to be some thing to which the “I” refers in the statement, “I think, therefore I am.” But I owe a better account of what sort of thing that thinking thing is—a thing that, in some way or another, strictly speaking, can be said to not exist. Or I need to concede that van Inwagen is right.
I believe my views can be filled in such a way that maintains a satisfying logical and metaphysical compatibility between compositional nihilism and the Cogito, just as I can do all sorts of things with chairs without believing that they exist, strictly speaking. Again, compositional nihilism doesn’t hold me to the view that “I” don’t exist, even strictly speaking, in the sense of being some sort of entity that has experience. My claim is simply that this entity is not a composite object, just as the chair is not a composite object.
In fact, the Cogito is just the thing for this predicament. There exists some experiencing thing, even if I can’t tell you quite what sort of thing that is. But I know it exists, because I experience. I can rightly say then that it experiences, therefore it exists. And, more to the point, as that experience is my own: I experience, therefore I exist.
At any rate, I suppose that the things I’ve said about the Cogito here today will need to hold up well against whatever I’ve said in the above-linked response to van Inwagen. Perhaps they hold up fine. But this is where I’ll leave things for today.
Enjoy or find this post useful? Please consider pitching in a dollar or three to help me do a better job of populating this website with worthwhile words and music. Let me know what you'd like to see more of while you're at it. Transaction handled by PayPal.