Parfit's On What Matters: Unifying Moral Philosophy

Imagine two people who agree that a promise should be kept. One points to the friend's reliance on it. The other asks what would happen if people generally abandoned commitments whenever something more attractive appeared. Agreement about the action is clear; agreement about its justification is less clear. A shared verdict can emerge from different reasons, and the difference may matter when the next case is harder.

Derek Parfit's On What Matters asks how far some influential approaches to morality can converge. Its first two volumes appeared in 2011; a third followed in 2017. The central convergence project concerns revised versions of Kantian ethics, rule consequentialism and contractualism. Parfit tries to show that their strongest formulations can support the same principles. Understanding that proposal requires more than checking whether three familiar theories happen to approve the same action.

The distinction changes how we read the book. Parfit is developing arguments for particular formulations, then arguments connecting them. He is not reporting that existing disagreements have disappeared. This chapter explains the proposed connections, the role of objective reasons, and the tests a reader should apply before accepting the resulting Triple Theory.

A reason, in the normative sense relevant here, is a consideration that counts in favor of an attitude or action. Someone's pain might count in favor of helping them. That differs from an explanation of why a person acted: fear of embarrassment could explain someone's assistance even when the other person's need supplies the better justification. The distinction lets us ask whether an action has good reasons behind it without assuming that its actual motive was good.

Parfit defends an objective account of reasons. On his view, what we have reason to do is not fixed simply by what we happen to want. Desires themselves can be assessed. A desire to avoid severe suffering can have a justification that an arbitrary desire to suffer on a particular day lacks. Describing a preference, even accurately, does not yet settle whether there is a reason to act on it.

This is an argument within the book, rather than a starting point every reader must accept. The question is whether facts about the objects of our desires can give us reasons independently of those desires. Parfit's opposition to desire-based accounts concerns the grounds of normative reasons; it does not require denying that desires influence behavior or that learning what someone wants matters in deliberation.

Consider a hypothetical person who refuses a painless way to avoid agony because they currently have no desire to avoid it. The example separates wanting relief from having reason to want relief. An objectivist can say that the suffering supplies a reason even when the desire is absent. Someone defending a desire-based view will need to explain which desires count and why. The example presents the dispute; describing it does not resolve every version of that dispute.

The objectivity question also differs from the convergence question. We could agree that there are objective moral truths while disagreeing about which theory describes them. Conversely, agreement among several theories would not establish objectivity without further argument. Parfit pursues both projects, but success in one cannot simply be assumed to establish the other.

Consequentialism explains moral assessment through consequences. For the convergence argument, the consequentialist component is a form of rule consequentialism. It evaluates principles through the consequences associated with their general or universal adoption, under the formulation being considered. Acts are then assessed through those principles. This differs from evaluating each individual act solely by whether that act produces the best available outcome.

The distinction matters for promises. Suppose that breaking a promise on one occasion would produce a small benefit. An act-based calculation and an assessment of a generally accepted principle concerning promise breaking ask different questions. The principle may affect assurance and reliance across many situations. A rule consequentialist must specify which consequences of accepting or following a principle count, rather than merely insert the act's immediate benefit into a larger-sounding description.

Kantian reasoning asks about principles that could serve as universal laws, and about the reasons everyone has for willing them. Parfit critically revises formulations associated with Kant. A reader should therefore distinguish what Kant originally wrote from the proposals Parfit thinks most defensible. His argument is a philosophical reconstruction, not a claim that every familiar interpretation of Kant already entails the Triple Theory.

Scanlonian contractualism evaluates principles through reasonable rejection. It concerns justification to individuals who seek terms of conduct others could also reasonably accept. Actual agreement is insufficient: someone may accept a bad arrangement because they lack power, information or confidence. The question concerns the quality of reasons for rejecting principles, not simply the number of people who object.

These approaches also place their tests at different points. An assessment of outcomes asks how things go. A test of rational willing asks what people have reason to choose as a universal law. Reasonable rejection asks what burdens or other objections individuals can press against a principle. Showing that these tests support the same principles requires a connecting argument. Renaming all three as concern for people would leave that work undone.

In the 28 April 2008 author draft listed below, Parfit's Triple Theory says that an act is wrong just when acts of that kind are disallowed by principles that are optimific, uniquely universally willable, and not reasonably rejectable. In this context, an optimific principle belongs to principles whose relevant universal adoption would make things go best. Universal willability concerns what everyone could rationally will as universal laws. Non-rejectability supplies the contractualist condition.

The proposal is about principles with these properties, rather than a vote among three theories. It should not be paraphrased as saying an act is wrong only when three independent moral judges all disapprove. Such a voting procedure could leave disagreements unresolved without explaining why the theories agree. Parfit wants to defend a common account of the principles that govern conduct.

Each condition brings a substantive question. What makes outcomes better? Which reasons may someone use in deciding what they could rationally will? What makes rejecting a principle reasonable? The compact formula does not answer these questions merely by containing three technical terms. Much of the book's argument concerns how to understand them and whether their requirements can be brought together.

The uniqueness qualification matters too. If several incompatible sets of principles could each be rationally willed, the proposed route from rational willing to a particular moral standard would need further work. Identifying a principle that someone could accept is weaker than showing that everyone has the relevant reasons to choose that principle over its alternatives. Readers should watch these changes in logical strength.

Parfit's discussion also distinguishes principles that are universally accepted from principles that are universally followed. Acceptance concerns how people regard and use a principle; following concerns conduct that conforms to it. These descriptions can come apart. A person might accept a prohibition yet violate it. An argument about the consequences of universal acceptance must therefore avoid silently substituting the consequences of universal compliance.

One important route begins with Parfit's Kantian Contractualist Formula. This concerns principles whose universal acceptance everyone could rationally will or choose. Parfit then argues for a connection between that test and principles whose universal acceptance makes things go best. If the argument succeeds, a Kantian approach supports a form of rule consequentialism.

A further step connects those principles to reasonable rejection. Parfit argues that the relationship can bring a Scanlonian condition into the account. The argument needs to establish whether optimific principles are the only ones whose universal acceptance everyone could rationally choose, and whether they are the only principles no one could reasonably reject. These are propositions to defend, not definitions that make convergence automatic.

An analogy can clarify the ambition without serving as proof. Imagine two different tests of whether a bridge design is acceptable. One assesses how its parts bear loads; another evaluates how the whole structure behaves. Agreement on a few designs would be encouraging, but a claim that the tests are equivalent would require explaining their relationship. Moral tests involve contested reasons rather than engineering measurements, so the analogy reaches only as far as that distinction between matching examples and an argument for equivalence.

The same discipline applies to counterexamples. If a proposed formulation gives an implausible result, Parfit can revise it, but the revision needs a reason beyond making the convergence argument work. A reader can ask whether the new formulation improves the theory on its own terms. Otherwise, the procedure risks selecting definitions to secure the desired agreement.

This is why a short account should not present the book as a demonstration that every conflict between duties and outcomes is illusory. The argument concerns selected, developed formulations. A person who rejects one reformulation may retain a coherent disagreement with the proposed synthesis. Understanding exactly where the disagreement remains is a useful result of reading the argument.

Return to a hypothetical promise. A friend has arranged transport around your offer of a lift. Keeping it is inconvenient, and you could instead spend the time on something enjoyable. We can examine the reasons under different principles without pretending that this invented case is a passage from Parfit.

An act-focused approach might begin by comparing the consequences of keeping and breaking this particular commitment. A rule-focused approach asks about principles permitting departures from assurances in circumstances of this kind. It considers how adopting those principles affects the ability to rely on commitments. The difference is one of what is being assessed, not simply whether the person doing the assessment cares about consequences.

A Kantian contractualist asks whether everyone could rationally will the relevant principle's universal acceptance. The friend whose plans depend on the assurance matters within that question, as does the person who made it. Stating their interests does not yet show what each could rationally will. We must specify the principle and the reasons available to both positions.

A Scanlonian approach asks whether a principle allowing cancellation for this kind of convenience could reasonably be rejected. The friend's reliance gives an objection to permission. The promiser's burden supplies a consideration against a stricter alternative. Comparing these considerations requires more than asking whether either person is unhappy with the result.

Agreement in this case would still leave harder cases. Suppose the promised lift conflicts with an urgent rescue. A defensible principle might distinguish inconvenience from an emergency. The work lies in defending that distinction and its limits. Adding an exception whenever our preferred verdict changes would conceal the question of principle that the example was meant to expose.

Two theories can agree about every example we have considered without providing the same explanation. One might treat reliance as valuable because of its beneficial consequences. Another might regard an individual's claim against betrayal as independently important. This difference can affect how each responds when reliance produces no further benefit or when honoring it has substantial costs.

Parfit's higher-level project does not require us to deny that particular acts have more specific wrong-making features. The draft formulation of the Triple Theory expressly allows other properties to make acts wrong and presents the combined condition as a higher-level account. The question is whether it can unify these features without erasing what explains their importance.

That distinction gives a more exact objection than saying the theories merely use different words. Different descriptions could identify the same morally significant structure. But it also gives a more exact standard of success than observing agreement: the proposed unification must explain why those descriptions fit together. Readers can assess whether the account preserves individual claims, rather than assuming that a common formula automatically does so.

Nor should we infer that failed convergence would make morality nothing more than preference. That would require a separate argument ruling out moral truth wherever disagreement persists. A failure to reconcile selected theories could show that one theory is mistaken, that their formulations need work, or that their relationship has been misunderstood. It would not by itself settle metaethics.

The first pressure point concerns the choice of formulations. Does a revision have support within the approach it is meant to improve? Someone drawn to Kantian ethics may resist a reformulation that appears to abandon the very constraint they wanted to defend. The proper response is to examine the reasons for the revision, rather than declare that any resistance is a misunderstanding.

A second concerns the best outcome. The word optimific can hide disagreement about the good. An account that includes distribution, suffering, achievement or other values must explain how they enter its comparisons. Calling a principle the best does not establish that everyone could rationally will it. The convergence argument needs the relevant conception of value and the relevant account of rational choice.

A third concerns individuals who bear severe burdens. A principle might improve outcomes overall while asking a particular person to accept a large loss. The route from that improvement to reasonable non-rejectability must address the person's objection. Simply pointing to the total benefit would assume a disputed answer to the contractualist question.

A fourth concerns idealization. Universal acceptance is a hypothetical condition, whereas actual moral decisions occur amid disagreement and incomplete information. We need to distinguish a criterion of right action from a procedure a person can use with limited knowledge. A theory might offer the former without making the latter easy. It would be misleading to sell the Triple Theory as a practical device that produces immediate consensus.

These are tests of the argument, rather than claims that a named critic has decisively refuted it. They help locate the reasoning that needs examination. A focused objection identifies a formulation, a premise or an inference. It gives us more to work with than a general impression that reconciling theories is either inspiring or impossible.

The first volume develops reasons, principles and the convergence project. The second contains critical discussions and Parfit's further arguments, including a chapter specifically devoted to the Triple Theory. The third extends the project. A first reading can focus on the relationship between rational willing and optimific principles, while keeping the larger defense of normative truth in view.

Choose one formulation and track what changes when Parfit revises it. Note whether the change concerns the object of evaluation, the information available to the chooser, or the reasons admitted into the choice. Then ask whether the connecting argument uses exactly that revised formulation. This exercise makes equivocation easier to detect without requiring a premature verdict on the whole book.

The reward is a more careful understanding of disagreement. Sometimes competing moral descriptions may support a common principle; sometimes their agreement conceals a remaining difference about justification. Parfit asks us to investigate that relationship through arguments we can challenge. The task is to establish how far the reasons converge, and to state precisely where they do not.

• Derek Parfit, On What Matters, author draft dated 28 April 2008. This draft supplies the Triple Theory formula and higher-level-property formulation discussed above; it precedes the published 2011 volumes listed next.
• Derek Parfit, On What Matters, Volume One. Begin with the treatment of reasons and follow the development of the convergence argument.
• Derek Parfit, On What Matters, Volume Two, chapter 23, “The Triple Theory”. This chapter examines the connections among optimific principles, rational choice and reasonable rejection.
• Derek Parfit, On What Matters, Volume Three. A later continuation of the project.
• T. M. Scanlon, What We Owe to Each Other. Reading its account of reasonable rejection alongside Parfit helps make the proposed reformulations visible.