Forgiveness

Philosophical and historical scholarship

The Stanford Encyclopedia of Philosophy states that interpersonal forgiveness is prominent in contemporary moral discourse but was neither examined nor encouraged by the ancient Greeks and was largely ignored in the history of Western philosophy, with Bishop Butler (1726) named as the most significant exception. The classicist David Konstan argues, in a book reviewed by Ilaria Ramelli, that forgiveness in the modern sense, with acknowledged wrongdoing, remorse and change of character, was absent from the Greek and Roman world and emerged only in the last three to four centuries.

Konstan's thesis is one scholar's account. The reviewer notes that he gives limited attention to medieval sources, and a historical claim about when forgiveness became a moral idea should be read with that caveat.

Religious and philosophical traditions

In Jewish tradition, Mishnah Yoma 8:9 states that Yom Kippur atones for transgressions between a person and God, but that for transgressions between a person and another it does not atone until the wrongdoer has appeased the other person. Ramelli's review reports that Konstan found the Hebrew Bible and the New Testament concerned mainly with divine forgiveness rather than interpersonal reconciliation, while the Mishnah and Talmud contain some interpersonal examples.

The Stanford entry says forgiveness is praised by the major world religions and that some Christians believe they have been commanded to forgive all who wrong them, while disagreement persists about when forgiveness is owed. A 2026 comparative article describes Buddhist forgiveness as letting go of anger and resentment to extinguish internal causes of suffering, and contrasts this with psychology's emphasis on relational repair. A 2025 interview study of 16 religious scholars in an Islamic setting reports that, for wrongs between people, seeking the victim's forgiveness is described as essential alongside divine pardon. These are single recent articles, not histories, and they do not date the traditions they describe.

Development as a psychological intervention

Akhtar and Barlow identify two process-based forgiveness interventions as the most widely used: Robert Enright's process model and Everett Worthington's REACH model, the latter attributed to Worthington (2001). They list three earlier meta-analyses, by Baskin and Enright (2004), Lundahl and colleagues (2008) and Wade and colleagues (2013). The sources reviewed do not give a dated account of when forgiveness interventions were first trialled.

Definitions in philosophy and psychology

The Stanford Encyclopedia describes paradigmatic forgiveness as a relation between a wrong action, a wrongdoer and a wronged victim, in which the victim moves on from the wrong while still judging it a culpable wrong. Many philosophers hold that it involves overcoming resentment, and the entry reports debate over whether all negative emotion must go or only hostile emotion.

In the psychological literature reviewed, Akhtar and Barlow define forgiveness as the decision to let go of negative resentment-based emotions, thoughts and behaviours and to develop positive regard, such as compassion, sympathy or pity, towards an offender. The philosophical and psychological definitions overlap but are not identical.

What forgiveness is distinguished from

The Stanford entry treats forgiving as incompatible with condoning, because the forgiver still judges the act wrong, and as different from excusing, because the wrongdoer is still held culpable. It states that it is possible to forgive without reconciling, that reconciliation can occur without forgiveness, and that merely forgetting is not forgiving.

Research labels and the self-directed form

Wade and colleagues limited their meta-analysis to interventions designed explicitly to promote forgiveness of a specific hurt and delivered by a professional, so 'forgiveness therapy' in that literature is narrower than general talk of forgiving. Self-forgiveness interventions form a separate research strand, and one 2024 review treated them separately from forgiveness of others. The sources reviewed supported no alternative name for the practice, and none used the phrase 'art of forgiveness'.

The Enright process model

Akhtar and Barlow describe the Enright model as having four phases. Uncovering involves identifying psychological defences, recognising and expressing anger over the offence and acknowledging its psychological harm. In the Decision phase, participants explore what forgiveness means, consider it as a response and commit to it. The Work phase involves cognitive reframing of the offender, developing empathy and compassion and accepting the pain. In the Deepening phase, participants are encouraged to find meaning in the suffering and to recognise that others are also hurt.

The REACH model and delivery formats

The REACH model is a five-step approach: recall the hurt, develop empathy for the offender, consider forgiveness an altruistic gift, commit to forgive, and hold on to forgiveness in times of difficulty. In the randomised trials pooled by Akhtar and Barlow, REACH was delivered only to groups. Wade and colleagues found that treatment dose and format varied across studies, and a 2014 trial tested a six-hour self-directed workbook adapted from REACH.

Facilitators and participants

The studies reviewed in these meta-analyses were delivered by professionals, in individual or group formats, to adolescents and adults who had experienced offences ranging from everyday interpersonal hurts to sexual abuse and civil conflict. A separate meta-analysis covers forgiveness education programmes for children and adolescents. The sources reviewed do not describe religious or ritual forgiveness practices, such as confession, as structured practices.

Outcomes that have been studied

Research has examined forgiveness interventions for forgiveness itself and for depression, anxiety, hope, anger and hostility, stress and distress, and positive affect in people who report a specific interpersonal hurt. A meta-analysis of programmes for children and adolescents examined forgiveness and anger. A 2024 review examined self-forgiveness interventions and other psychological outcomes.

Claims made for the practice

Authors of a 2026 scoping review describe forgiveness interventions as having emerged as an effective treatment for trauma survivors. That framing is the authors' own, and the same review reports contrary findings in some contexts, discussed under safety. A 2001 survey of 134 survivors of human rights abuses in South Africa found low forgiveness attitudes associated with poorer psychiatric status. This was an association in one survey, not evidence that raising forgiveness improves health.

The sources reviewed do not test popular claims that forgiveness is required for recovery, or that it improves physical health. The outcome measures in these reviews were mental-health and forgiveness measures only.

Meta-analysis of professionally delivered interventions

Wade and colleagues pooled 53 post-treatment effect sizes (2,323 participants) and 41 follow-up effect sizes (1,716) from 54 reports. Participants receiving explicit forgiveness treatment reported more forgiveness than untreated participants (Delta+ 0.56, 95% confidence interval 0.43 to 0.68) and than participants receiving alternative treatments (0.45, 0.21 to 0.69). Depression, anxiety and hope improved more than with no treatment. Longer interventions and individual delivery predicted larger gains in forgiveness.

The abstract does not state whether effects persisted at follow-up, and the authors note that no individually delivered studies used the Worthington model, so differences between models and formats cannot be cleanly separated.

Meta-analysis of randomised trials

Akhtar and Barlow reviewed 15 randomised trials against wait-list or no-treatment controls and meta-analysed 12. They reported lower depression (standardised mean difference -0.37, 95% CI -0.68 to -0.07), anger and hostility (-0.49), and stress and distress (-0.66), and higher positive affect (-0.29), with the signs following the review's scale coding. Pooled state forgiveness effects across 11 studies were small to moderate (-0.43) with substantial heterogeneity.

The authors called the evidence moderately strong but rated the quality of included studies low to moderate, mainly because blinding is hard in this field. Comparators were wait-list or no-treatment, so the review does not show forgiveness work to be better than other therapies. The Enright and REACH subgroups were small (3 and 7 studies) and confounded by format.

Other designs and populations

A 2022 meta-analysis of 20 studies with 1,472 young people found forgiveness education interventions had a significant effect on forgiveness (Hedges g 0.54) and anger (g 0.29). A 2024 systematic review of 21 articles on self-forgiveness interventions found varied methods and outcomes and reported that Enright-model self-directed interventions appeared effective. A 2014 randomised trial of a REACH-based workbook, in 41 undergraduates against a wait-list, reported increased forgiveness.

Pressure to forgive and abusive situations

A 2026 scoping review of 54 studies on forgiveness and adverse childhood experiences found that self-forgiveness was largely linked to better outcomes, that results for forgiving others were mixed, and that survivors with severe abuse histories typically found forgiving harder. It reported harmful effects when survivors felt pressured to forgive, or pressured not to forgive, and concluded that forgiveness work needs trauma-informed delivery that prioritises autonomy and safety.

The Stanford entry notes philosophers' concern that forgiving a dangerous, unrepentant wrongdoer may facilitate reoffence in some circumstances, and it treats forgiveness as separable from reconciliation, so forgiving does not entail resuming contact. It reports that philosophers disagree about whether forgiveness is always good.

Adverse effects and evidence gaps

The meta-analyses reviewed for this draft do not report systematic adverse-event data, and their trials were mostly short with wait-list comparators. Absence of reported harm is not evidence that none occurs. Findings from consenting participants in structured programmes should not be transferred to situations in which a person is being urged to forgive someone who remains a danger.

Training and care boundaries

No statutory regulation or practitioner standard specific to forgiveness work was found in the sources reviewed. The International Forgiveness Institute describes its own continuing-education course on forgiveness therapy, based on a manual by Robert Enright and Richard Fitzgibbons, for licensed psychologists and other professionals, and a certificate of completion. Its page does not describe licensing, and the certificate is a statement by the organisation itself.

Forgiveness work is not a replacement for safety planning or for assessment and treatment of depression, trauma or other conditions. Anyone in danger from a person who has harmed them, or experiencing severe distress, should seek appropriate professional or emergency help.