This is the teaching with the largest real-world test behind it, and the test is more contested than the admiring accounts allow.
Retributive justice treats a wrong as a debt: harm was done, harm is owed, the balance is restored by inflicting it.
Restorative justice treats a wrong as a tear: something between people has broken, and the question is what would repair it. Ubuntu is the philosophical justification usually given for the second, and it was given at enormous scale — the Truth and Reconciliation Commission, which handled South Africa’s transition out of apartheid through public testimony and conditional amnesty rather than trials.
Tutu, who chaired it, made the argument explicitly in ubuntu terms: that the humanity of the perpetrator and the victim are bound together, and that a justice which destroys the wrongdoer does not repair the whole.
That is a serious argument made under maximum pressure about the worst possible material, which is why it deserves better than to be quoted approvingly and moved past.
The Commission is frequently invoked in Western writing as a demonstration that forgiveness works. The South African assessment is considerably more divided, and the criticisms are not marginal.
Amnesty was granted to people who had tortured and killed, in exchange for full disclosure. Many victims’ families experienced that as their claim being taken from them without their consent — the state trading their case for national stability. The Biko family challenged the amnesty provisions in court on exactly those grounds.
Reparations were recommended and then delivered slowly, partially and at a fraction of what was proposed. People testified about the worst thing that had happened to them and received very little.
And there is a structural critique made at length by South African scholars: that the framework required the dispossessed to perform forgiveness for the benefit of a settlement that left the economic structure of apartheid largely intact. Reconciliation without redistribution.
Some of this is an argument about implementation rather than about the philosophy. But an essay that uses the TRC as evidence for restoration has to carry the whole verdict, and the whole verdict is mixed.
Away from the transitional-justice case, the research on restorative approaches is reasonably encouraging and reasonably bounded.
Restorative conferencing in criminal justice shows modest reductions in repeat offending compared with conventional processing in several trials, and consistently higher satisfaction among victims — which is the finding that surprised people, since the assumption was that victims wanted punishment.7
The forgiveness literature is largely about the person forgiving: letting go of a grievance is associated with better outcomes for them.5 That is worth separating carefully from the justice question. It is an argument about what carrying a grudge does to you, not an argument about what anyone is owed.
And at the scale of a relationship, the repair attempt — the move that interrupts an escalating conflict — is among the better predictors of whether a relationship survives.6
Restorative language is used to pressure people into premature reconciliation more often than it is used to achieve repair.
The pattern is consistent: the person with less power is asked to be the bigger person, to let it go, to think of the group. Institutions reach for restoration specifically when accountability would be expensive. “We are all one here” has been said to a great many people immediately before they were asked to accept something.
The tradition is not naive about this — the TRC required full disclosure, and restorative processes generally require the wrongdoer to acknowledge what they did. That requirement is the whole safeguard, and it is the first thing dropped when the framework is borrowed.
So: repair is the better question, and forgiveness belongs to the person wronged and nobody else. Anyone else recommending it, particularly anyone with a stake in the matter being closed, is doing something other than philosophy.
Note — who is asked to be the bigger person
Restorative processes require the wrongdoer to acknowledge what they did — that requirement is the safeguard, and it is the first thing dropped when the framework is borrowed. Forgiveness belongs to the person wronged. Anyone with a stake in the matter being closed is not offering philosophy.
At ordinary scale, where most of this is actually available.
Ubuntu is a living idea belonging to living peoples, not a closed corpus, and the terms used here come from several distinct languages — Nguni (isiZulu, isiXhosa), Sotho-Tswana and others. Where a word is an ordinary word in its language rather than a technical term, the essay says so; where a phrase’s provenance is uncertain or its popular gloss is a later Western elaboration, the essay says that too. African scholarship on ubuntu is cited in preference to Western commentary wherever it bears on the point. Links are given where a stable public page exists.
Restoration asks the better question — what would repair this, rather than what does this person owe. The requirement that the wrongdoer acknowledge what they did is the whole safeguard, and the first thing dropped when the idea is borrowed.
And the TRC, the case always cited for it, is far more contested in South Africa than in the accounts that quote it: amnesty without consent, reparations largely undelivered, reconciliation without redistribution.