Why you can't force yourself to intend the impossible
A billionaire offers you one million dollars. All you have to do is intend, at midnight tonight, to drink a painful but non-lethal toxin tomorrow afternoon. You don't actually have to drink it; the money is paid tomorrow morning based on a brain scan of your sincere intention tonight. Can you do it? Philosopher Gregory Kavka argued you cannot sincerely intend to do an action you know you will have no reason to perform.
The Billionaire's Offer
Imagine an eccentric billionaire approaches you with a peculiar proposition. He places a small vial of clear liquid on the table. The liquid is a mild toxin: drinking it will cause several hours of intense nausea, fever, and acute discomfort, but it carries no lasting damage, no risk of death, and requires no hospitalization. Beside the vial, he places a check for one million dollars.
The conditions of the deal are unusually specific. You do not actually have to drink the toxin to get the money. Instead, the billionaire possesses an infallible brain scanner capable of reading your mental state. At midnight tonight, the scanner will assess whether you sincerely intend to drink the toxin tomorrow afternoon at three o'clock. If the scanner detects a genuine, unforced intention to drink the poison, the one million dollars will be wired directly into your bank account tomorrow morning at eight o'clock.
At eight o'clock tomorrow morning, you check your account. The money is either there or it is not. Come three o'clock tomorrow afternoon, the vial sits before you. Because the transaction was completed hours earlier, drinking the poison will not add a cent to your wealth, and pouring it down the drain will not take a cent away. Knowing this, the question is simple: can you successfully form the sincere intention at midnight?
The Gap Between Desiring and Intending
First formulated by the American philosopher Gregory Kavka in 1983, the toxin puzzle exposes a fundamental fault line in how we understand human rationality, choice, and agency. At first glance, winning the money seems trivial. You simply tell yourself, 'I will drink the toxin tomorrow.' You desperately want the million dollars, and you know that having the intention is the sole key to unlocking the fortune. Why would anyone fail such an easy mental exercise?
The problem arises because wanting to have an intention is not the same thing as actually having one. You can easily desire to possess the intention to drink the toxin because possessing that intention yields a massive reward. However, an intention is an active commitment to carry out a specific future action. In normal human psychology, rational agents form intentions to act based on the reasons favoring the action itself, not based on the independent benefits of merely holding the mental state.
If you ask yourself why you should drink the poison tomorrow at three o'clock, you realize there is zero reason to do so. The payout has already occurred or failed based on your state at midnight. Drinking the liquid brings only agony with no upside. Because you know with certainty that you will have no reason to drink the toxin when the moment comes, your rational mind cannot spontaneously generate a genuine intention to do so in advance.
Why You Cannot Trick Yourself
Faced with this barrier, many people immediately look for psychological escape hatches. A common initial reaction is to think, 'I will just decide to drink it, and if I change my mind tomorrow afternoon, that is fine.' But Kavka points out that this strategy collapses under scrutiny. If your midnight plan includes the implicit clause that you will abandon the action once the money arrives, you never held an intention to drink the toxin in the first place; you merely held an intention to feign the commitment.
An intention requires a settled resolution to act. If a rational agent knows that an upcoming decision point will present an overwhelming reason to choose Option B over Option A, and no reason whatever to choose Option A, they cannot sincerely resolve in advance to choose Option A without self-deception. Since the scanner is defined as a perfect detector of genuine mental states, any superficial mental performance or self-delusion will register as a lack of genuine intent.
The puzzle reveals a stubborn asymmetry in rational decision-making. You have an overwhelming reason to intend to drink the toxin (one million dollars), but you have no reason at all to actually drink the toxin. Kavka demonstrated that reasons for intending do not automatically translate into reasons for acting, creating a scenario where a perfectly rational person is barred from obtaining a reward precisely because of their rationality.
Failed Workarounds and External Devices
Another natural impulse is to introduce external constraints to force compliance. For example, what if you write a legally binding contract promising to donate ten million dollars to an enemy if you fail to drink the toxin, or what if you take a drug at midnight that temporarily suspends your ability to back out tomorrow? While these strategies might successfully compel you to drink the poison, they sidestep the philosophical challenge rather than solving it.
Kavka's thought experiment explicitly tests internal mental agency—whether a person, using reason alone, can adopt a pure intention. Introducing external stakes simply changes the payoffs so that drinking the toxin becomes locally rational tomorrow afternoon to avoid an even worse penalty. In doing so, you have not solved the problem of intending an unmotivated act; you have merely created a new, artificial motive for the act itself.
If the rules forbid changing your future payoffs or altering your brain chemistry, you are left with only your unassisted willpower. In that state, you sit in bed at midnight staring at the clock, acutely aware that tomorrow you will have every reason to pour the vial down the sink. That unavoidable future reality echoes backward in time, dissolving your ability to make a sincere promise to yourself in the present.
The Shadow of Nuclear Deterrence
The toxin puzzle was not conceived merely as an entertaining parlor game. Gregory Kavka developed the paradox while examining the grim logic of nuclear deterrence during the Cold War. The policy of Mutually Assured Destruction required nations to possess a credible, sincere intention to launch a devastating retaliatory nuclear strike if attacked by an adversary.
The strategic paradox of deterrence mirrors the toxin puzzle directly. Having the intention to retaliate serves a vital purpose beforehand: it deters the enemy from launching a first strike, potentially preserving peace. However, if deterrence fails and the enemy's missiles are already in the air, actually carrying out a retaliatory strike achieves nothing positive. It does not undo the incoming destruction; it merely ensures the end of civilization. At the moment of execution, launching retaliation becomes an act with catastrophic costs and no rational upside.
Kavka used this connection to show that maintaining a credible deterrent requires human leaders to maintain an intention that would be utterly irrational to execute if the time ever came. Just as with the toxin, the immense value of having the intention exists entirely before the action, whereas executing the action itself remains completely ungrounded in constructive reasons.
What the Paradox Teaches About Rationality
Kavka's paradox prompted extensive debate in decision theory and the philosophy of mind. Philosophers such as David Gauthier and Michael Bratman have used the puzzle to explore how rational agents coordinate their actions across time. Traditional rational choice theory evaluates decisions strictly based on the expected utility at the precise moment of choice. Under this standard act-based view, an agent evaluates only what lies ahead at 3:00 PM, making refusal inevitable.
Some philosophers argue that this standard view is too narrow. They propose that rationality ought to allow for broader plans or 'constrained maximization,' where an agent adopts a disposition to follow through on commitments because being that type of person yields superior overall life outcomes. If your framework of rationality prevents you from walking away with a million dollars from a harmless afternoon of nausea, critics argue that standard rationality may itself be flawed.
Ultimately, the toxin puzzle endures because it demonstrates the limits of voluntary control over our own beliefs and intentions. We cannot choose what to intend purely because it is profitable to intend it, any more than we can genuinely force ourselves to believe the sky is green in exchange for a cash reward. Our intentions remain inextricably tethered to the reasons we see for acting, exposing the deep boundaries between our desires and our will.
Key takeaways
•Gregory Kavka's toxin puzzle shows that having an incentive to intend an action is fundamentally different from having an incentive to perform that action.
•A rational agent cannot sincerely intend to perform an act if they know that, when the time comes to act, they will have zero reason to carry it out.
•The thought experiment originated from Cold War dilemmas in nuclear deterrence, where maintaining a sincere threat is useful in advance even if executing the threat would be pointless after an attack.
•The paradox challenges traditional rational choice theory by revealing scenarios where pure, forward-looking rationality prevents people from achieving the best possible outcome.