Symposium · Ethics · Pl. 10Immanuel Kant · 1724–1804
Humanity as an End in Itself
A second angle on the same rule
The categorical imperative asks whether a rule could be willed as a universal law. Kant did not think that was the only way to state the same underlying moral demand. He offered a second formula, aimed more directly at how people actually treat one another, that many find easier to apply in the moment than the abstract universalizing test. You do not need to imagine a world of liars to see what is wrong with the lie you are about to tell. You need only look at the person in front of you and ask whether, in telling it, you have made them into a tool. Kant thought that question and the universal-law question had the same answer, every time.
The world it came from
Kant spent his whole life in Königsberg, a Baltic port in Prussia, and never travelled more than a few miles from it. He was raised in a strict Pietist household, taught at the university, and published the Groundwork of the Metaphysic of Morals in 1785, four years after the Critique of Pure Reason. The moral philosophy of his day derived duty from something else: from feeling, as Hume had, from God's command, or from the pursuit of happiness. Kant thought every such derivation made morality hostage to a fact that could change. Rousseau, whose portrait was the only one on his wall, had taught him that the common person's sense of duty deserved respect. The Groundwork set out to show what that sense already contained.
The idea
Treat humanity, whether in your own person or in that of any other, always as an end, and never merely as a means. Using someone purely as an instrument for your own purposes, lying to get their help, manipulating their trust, ignoring their own goals entirely, violates this formula even when the universalizing test might not obviously catch it. Rational beings are owed a status other tools are not. Kant's word for that status is dignity, which he set against price. Anything with a price can be exchanged for something else of equal value. A person cannot, because a person is the one who does the valuing. Humanity, for Kant, means that capacity to set ends, and it is the only thing that is good without condition.
How the reasoning runs
Start with what an imperative needs: an end that gives it force. Second, note that every end you happen to want, comfort, wealth, a finished project, has only relative value; it matters because you want it, and a different person might not. Such ends can ground only conditional rules. Third, a categorical rule would need an end whose value does not depend on anyone wanting it. Fourth, Kant claims there is exactly one candidate: rational nature itself, the capacity to set ends at all. Every relative value presupposes it, since nothing has value until a rational being confers it. Therefore rational nature is an end in itself, and the rule follows. Never treat the source of all value as if it were merely one more thing to be used.
Seen in life
You need a colleague to cover your shift on Saturday. You could tell her the truth: you have tickets to a match. You could tell her your mother is unwell, which will work faster and cost you nothing. Run Kant's second test. The lie does not merely mislead her. It removes her ability to decide, because you have arranged the facts so that the decision she thinks she is making is not the one she is making. She has become a lever you pull. The truth might get a no. But a no from a person who knew what she was agreeing to is a no you were owed. Kant's formula bans the shortcut precisely because it works.
Why it matters
This gives moral reasoning something the universal-law formula alone can feel thin on: a direct account of what it is to wrong a particular person, not just what happens if everyone adopted your rule. Kant believed the two formulas ultimately point at the same underlying law, but this version names the harm in manipulation and exploitation in language nearly anyone can immediately feel. It also draws a line that consequences cannot cross. If a person has dignity rather than price, then no sum of benefits to others can license using her up. That thought sits underneath the modern language of human rights, and underneath the rule that a patient must consent to what is done to her body, whoever it would help.
“So act as to treat humanity, whether in thine own person or in that of any other, in every case as an end withal, never as means only.”
From the source
Kant states the formula directly, right after deriving it from the universal-law version in the Groundwork: “So act as to treat humanity, whether in thine own person or in that of any other, in every case as an end withal, never as means only.” He then runs it through the same four examples he had used for the first formula. The person who makes a false promise, he says, is using the other as a mere means, since the promisee cannot possibly agree to the way he is being treated. The person who neglects his talents fails to treat the humanity in himself as an end. And the person who never helps anyone agrees with humanity as an end only negatively, not, as Kant puts it, positively.
The pushback
A common objection: doesn't ordinary life require treating people partly as means constantly, a cashier processing your purchase, a doctor treating a patient, without any of that being wrong? Kant's own qualifier answers this directly: the rule bans treating someone as a means only, not treating them instrumentally at all, as long as their own rational agency and consent are also respected throughout. A harder objection is that the qualifier does all the work and Kant never defines it. When is consent enough? A worker who accepts a dangerous job because the alternative is hunger has consented. So has a patient who signs a form she did not read. If the formula cannot say which agreements count, it cannot say who is being used.
The reply
Kant's own test for consent is whether the other person could share your end: could the promisee, told the truth, agree to the deal as you actually intend it? That rules out deception and coercion. It leaves the hungry worker unresolved, and Kant's later followers have spent much effort here. Onora O'Neill argued that the standard is possible consent, not actual: could the person have refused? Christine Korsgaard added that the formula obliges you to make refusal possible, which pushes toward fair conditions rather than mere signatures. These are developments, not answers Kant gave. The formula tells you what using someone is; it does not tell you, alone, how much a fair society must fix before agreement is real.
Where it went next
The word dignity travelled far. It opens the German constitution of 1949, whose first article declares human dignity inviolable, and it stands in the preamble of the Universal Declaration of Human Rights. In medicine the formula became the doctrine of informed consent, and the Nuremberg Code of 1947 reads like an application of it. Rawls built A Theory of Justice on a self-described Kantian foundation. Korsgaard and O'Neill rebuilt the argument for a new generation, and Derek Parfit's On What Matters in 2011 took the formula apart clause by clause and tried to reconcile it with consequentialism. Critics of factory labor, of surveillance, and of the attention economy reach for the same phrase, often without knowing they are quoting a Prussian from 1785.
Where this leads
The trolley problem is a natural stress test for this exact formula, asking whether pushing a stranger off a bridge uses him merely as a means in a way that steering a runaway trolley toward him does not, even though both actions end in his death. The categorical imperative is the first formula, the one this beat restates; Kant insisted the two are the same law seen from different sides, and readers have argued ever since about whether that is true. One word for a whole life, from Confucius, is the oldest cousin: do not do to others what you would not want done to you already asks you to see the other person as a self, which is most of what Kant meant by an end.