Acceptance and Commitment Therapy

Abstract

Acceptance and Commitment Therapy is a contextual form of cognitive behavioral therapy, itself a kind of behavior therapy, that treats psychological suffering as the product of rigid, avoidant responses to inner experience rather than of the inner experience itself. Developed by Steven Hayes and colleagues from the 1980s and grounded in relational frame theory, it aims not to reduce symptoms directly but to raise psychological flexibility: the capacity to stay in contact with the present moment and act on chosen values while accepting unwanted thoughts and feelings. This article sets out its six-process flexibility model, the relational-frame account of why language generates distress, the experiential-avoidance paradox at its core, and a meta-analytic evidence base whose strength depends heavily on the comparison condition against which the therapy is tested.

Keywords: Acceptance and Commitment Therapy, psychological flexibility, experiential avoidance, relational frame theory

Acceptance and Commitment Therapy, usually spoken as one word rather than as initials, belongs to what its originator called the third wave of behavior therapy: a group of contextual approaches that followed classical behavior therapy and cognitive therapy and that share a shift of target from the content of thought to a person's relationship with it (Hayes, 2004). Where cognitive therapy works to change the accuracy of dysfunctional thoughts, ACT leaves their content alone and works instead to change their function, loosening the grip that thoughts and feelings exert over action (Hayes et al., 2006). The therapy rests on a specific theory of language and cognition, and it defines its goal not as symptom relief but as a broadened, more flexible behavioral repertoire in the service of what a person values.

What Acceptance and Commitment Therapy Is

ACT is a form of psychotherapy within the contextual behavioral tradition, and MeSH classifies it under cognitive behavioral therapy, a branch of behavior therapy. It shares with that family an empirical, learning-based orientation and a concern with observable action, but it parts company with standard cognitive therapy on a basic point. Cognitive therapy holds that distorted thoughts cause distress and that correcting them relieves it; ACT holds that the struggle to control unwanted thoughts and feelings is itself the more remediable source of suffering, and that the aim of treatment is to reduce that struggle rather than the thoughts (Hayes et al., 2006).

The name states the two halves of the method. Acceptance is the active, willing contact with private experiences — sensations, emotions, memories, urges — without trying to alter their frequency or form, especially when doing so would cost behavioral effectiveness. Commitment is sustained action in the direction of chosen values, action maintained even in the presence of the unwanted experiences that acceptance makes room for. The therapy is thus not a relaxation or feel-better technique; a person practicing it may feel considerable distress while nonetheless acting in a way that a valued life requires (Hayes et al., 1999).

Because its stated end is not the removal of symptoms, ACT is often described as targeting a transdiagnostic process rather than a disorder. The same core problem — a rigid, avoidance-dominated relationship to inner experience — is held to run beneath many presentations, from anxiety and depression to chronic pain and substance use, which is why the model is applied across diagnostic categories with the same processes in view (Hayes et al., 2011).

The Psychological Flexibility Model

The organizing construct of ACT is psychological flexibility: contacting the present moment fully and as a conscious being, and persisting in or changing behavior in the service of chosen values (Hayes et al., 2006). The model decomposes flexibility into six processes, conventionally drawn at the points of a hexagon and nicknamed the hexaflex. The six are not independent techniques but facets of a single capacity, each paired with a form of rigidity that defines the corresponding pole of psychological inflexibility (Kashdan & Rottenberg, 2010).

The six group into three response styles. An open style comprises acceptance, the willing embrace of inner experience, and cognitive defusion, the ability to notice thoughts as thoughts rather than as literal truths. A centered style comprises present-moment contact, flexible attention to the here and now, and self-as-context, the experience of an observing self distinct from the thoughts and feelings it observes. An engaged style comprises values, freely chosen life directions, and committed action, patterns of effective behavior built in their service (Hayes et al., 2006). Psychological inflexibility is the mirror image: experiential avoidance, cognitive fusion, a rigid attention captured by a conceptualized past and future, attachment to a conceptualized self, unclear or unheld values, and inaction or impulsive, avoidant persistence.

Two features of the model matter for how the therapy is delivered. First, the processes are positive targets to build, not deficits to remove, so treatment develops willingness and valued action rather than eliminating symptoms. Second, because the six are interdependent, work on one — say, defusing from a harsh self-judgment — is expected to move the others, which is why ACT protocols mix experiential exercises, metaphors, and mindfulness practices rather than proceeding one module at a time (Hayes et al., 1999).

Relational Frame Theory and Language

ACT is unusual among therapies in resting on an explicit basic-science account of language and cognition, relational frame theory. The theory holds that the core of human language is the learned ability to relate events mutually and in combination — to derive, without direct training, that if A is like B then B is like A, and that if A is more than B and B is more than C then A is more than C. These derived relational responses, organized into relational frames, let arbitrary stimuli acquire the functions of the events they are related to (Hayes, Barnes-Holmes, & Roche, 2001).

The clinical consequence is that language allows suffering to be generated in the absence of any external cause. A word, a memory, or a private image can acquire the aversive functions of the event it frames, so a person can be frightened, shamed, or defeated by their own cognition with no threat present. This is the process ACT calls cognitive fusion: the tendency to respond to thoughts as though they were the things they refer to, so that the thought I am worthless functions as worthlessness itself rather than as a passing verbal event (Hayes, 2004). Relational frame theory thereby explains why simply disputing a thought's content, as in cognitive restructuring, may be limited: challenging a thought is still to engage its verbal network, whereas defusion changes the context in which the thought is held so that its literal meaning loses behavioral force (Hayes et al., 2006).

The same analysis grounds the therapy's approach to values. If verbal relations can attach aversive functions to neutral events, they can equally attach appetitive, motivating functions to chosen directions, letting a stated value organize behavior over long spans and in the face of immediate discomfort. Values work in ACT is the constructive use of the very relational capacity that, turned against a person, produces fusion and avoidance (Hayes, Barnes-Holmes, & Roche, 2001).

Experiential Avoidance

If one process is central to the ACT account of psychopathology it is experiential avoidance: the attempt to alter the form, frequency, or situational sensitivity of unwanted private experiences, even when doing so causes behavioral harm (Hayes et al., 1996). Avoidance is not irrational in the short term — suppressing a feeling, escaping a situation, or numbing a memory often brings immediate relief — and it is precisely that reliable short-term payoff that makes the strategy so persistent. The ACT claim is that as a general life strategy it backfires, because the effort to not feel narrows behavior, consumes attention, and paradoxically increases the very experiences it targets (Hayes et al., 1996). The paradox is not merely asserted. Classic experiments on deliberate thought suppression found that instructing people not to think of a target thought produces a rebound in which the thought returns more often once the effort relaxes, an ironic effect of trying not to think something that gives the ACT critique of experiential control a direct empirical anchor (Wegner et al., 1987).

The construct is transdiagnostic by design. Avoidant repertoires appear as the worry and reassurance-seeking of anxiety, the withdrawal of depression, the numbing of trauma, and the consumption of substance use, so a wide range of presentations can be read as variations on the same functional theme (Kashdan & Rottenberg, 2010). Because the harm is functional rather than topographical, ACT does not ask whether a behavior looks pathological but whether it works — whether it serves valued living or only the short-term regulation of discomfort.

Experiential avoidance is also the model's most measured construct. The Acceptance and Action Questionnaire and its revision, the AAQ-II, provide a self-report index of psychological inflexibility and experiential avoidance that has anchored much of the empirical literature, letting studies test whether change in avoidance mediates the outcomes ACT produces (Bond et al., 2011). That measurement work is what allows the avoidance account to be evaluated rather than merely asserted, though the questionnaire's overlap with general distress remains a live methodological question.

The Evidence Base

ACT has accumulated a substantial trial literature, and successive meta-analyses have reached broadly convergent but carefully qualified conclusions. Early quantitative reviews found medium-sized effects favoring ACT over control conditions across a range of problems (Powers et al., 2009), and a later meta-analysis of clinically relevant mental and physical health problems similarly reported that ACT outperformed control conditions and performed comparably to established treatments (A-Tjak et al., 2015). Reviews focused on anxiety and depression have concluded that ACT is an effective, evidence-supported option for both (Twohig & Levin, 2017); (Bai et al., 2020).

The size of the effect, however, depends heavily on what ACT is compared against, and this is where the literature is most contested. A methodologically stringent review argued that many ACT trials suffered from small samples, weak control conditions, and limited blinding, and that when these are taken into account the evidence supports ACT as probably efficacious for some conditions but does not establish superiority over other active treatments (Öst, 2014). A review of meta-analyses reached a similar overall reading: the empirical status of ACT is solid against inactive comparators but modest and uncertain against active ones, and the field's methodological quality, while improving, still limits strong claims (Gloster et al., 2020). Group-format delivery shows the same pattern of benefit over waitlist with less clear advantage over active care (Coto-Lesmes et al., 2020).

A second strand of evidence concerns not whether ACT works but why. Component and laboratory studies that isolate single processes — a brief defusion or acceptance manipulation, for instance — find that these components move the outcomes the flexibility model predicts, supporting the claim that the therapy operates through the processes it names rather than through nonspecific factors alone (Levin et al., 2012). This process-level evidence is central to the ACT research program, because a therapy defined by its mechanism must show that the mechanism, and not merely the package, is doing the work.

Acceptance and Commitment Therapy in Motion

The three demonstrations below make the model manipulable. The first builds a psychological-flexibility profile from the six processes, showing how each flexible pole and its inflexible opposite combine into an overall index. The second runs the experiential-avoidance paradox, showing how a strategy that pays off immediately erodes valued living as it is relied on over time. The third varies the comparison condition in an efficacy estimate, showing why the same therapy can look strong or weak depending only on what it is tested against.

Flexibility index 3 of 6 — a mixed profile. Open 1/2, centered 2/2, engaged 0/2.

Each process sits at either its flexible pole or its inflexible opposite; the index simply counts the flexible poles. The three style totals show that flexibility is a coordinated set rather than one skill — a person can be open yet disengaged, or committed yet fused — so it is the profile, not a single number, that names where therapy would work.

The flexibility demonstration assembles a profile across the six processes. Toggling each process between its flexible pole and its inflexible opposite — acceptance against avoidance, defusion against fusion, and so on — builds a psychological-flexibility index and shows the paired response styles filling in. It makes concrete that flexibility is not a single skill but a coordinated set, and that a person can be open yet disengaged, or committed yet fused, so the profile rather than a single number is what the therapy addresses.

At avoidance 0.80 over 10 periods, avoidance buys 8.0 units of comfort but leaves only 6.0 units of valued living.

The comfort avoidance buys rises in a straight line, which is why the strategy is so durable; the valued living it costs falls faster the more it is relied on and the longer the horizon, because each avoidant period narrows the accessible life by a compounding factor. That gap between immediate payoff and long-run cost is the paradox ACT works to expose.

The avoidance demonstration runs the paradox at the center of the model. Setting how heavily a person relies on experiential avoidance, and over how long a horizon, computes both the immediate comfort avoidance buys and the valued living it costs as its restricting effect compounds. It shows why avoidance is so durable — the comfort is real and immediate — and why ACT nonetheless treats it as the problem: the comfort accrues in a straight line while the cost to a life grows faster the longer avoidance is relied on.

Against inactive waitlist / no treatment, d = 0.8, so the average treated patient does better than 79% of control patients.

The therapy does not change between these rows; only the comparison does. A large, impressive effect against an inactive waitlist shrinks toward a coin toss against an established active treatment, which is exactly the interpretive issue that divides the ACT outcome literature: an effect size is meaningless until its comparator is named.

The efficacy demonstration makes the comparator problem visible. Choosing the condition ACT is tested against — an inactive waitlist, a nonspecific control, or an established active treatment — sets a standardized effect size and translates it into the share of control patients the average treated patient exceeds. It shows how a headline claim of effectiveness can shrink toward zero as the comparison is made more stringent, which is exactly the interpretive issue that divides the ACT outcome literature.

Worked Example

Begin with the flexibility index the first demonstration builds. Each of the six processes contributes one point when it sits at its flexible pole and zero at its inflexible opposite. A person who is accepting, defused, present, self-as-context, values-clear, and committed scores 1 + 1 + 1 + 1 + 1 + 1 = 6, full flexibility. A person dominated by avoidance, fusion, distraction, self-attachment, unclear values, and inaction scores 0. A mixed profile — open and present but fused and disengaged, say, accepting, present, and self-as-context flexible but defusion, values, and committed action inflexible — scores 3 of 6, and the point of the profile is that these three intact processes and three rigid ones name exactly where the therapy would work, which a single summed number would hide.

Now the avoidance paradox. Model reliance on avoidance as a value a between 0 and 1 over a horizon of T periods. Each period, avoidance buys immediate comfort equal to a, so cumulative comfort rises linearly to a times T. But chronic avoidance narrows the accessible life-space by a constant factor each period, so that space shrinks geometrically as r to the power t, with r = 1 − 0.15a; cumulative valued living is the geometric sum (1 − r to the power T) / (1 − r). Take a heavy avoider at a = 0.8 over T = 10 periods: r = 0.88, cumulative comfort = 0.8 × 10 = 8.0, cumulative valued living = (1 − 0.88^10) / 0.12 = 6.01. A light avoider at a = 0.2 gets r = 0.97, comfort = 2.0, but valued living = (1 − 0.97^10) / 0.03 = 8.75. The heavy avoider has bought four times the comfort and lost roughly a third of the valued living — the paradox in numbers.

Finally the comparator problem. Standardized effect size (Cohen's d) can be read as the fraction of control-group patients the average treated patient exceeds, U3 = Φ(d), the normal cumulative distribution at d. Against an inactive waitlist an illustrative d = 0.8 gives Φ(0.8) = 0.788, so the average ACT patient does better than about 79% of controls — a large, impressive-looking effect. Against an established active treatment an illustrative d = 0.2 gives Φ(0.2) = 0.579, only 58%, a difference that could as easily reflect residual bias as real superiority. The therapy has not changed between the two rows; only the comparison has, which is why an effect size is uninterpretable until its comparator is named (Öst, 2014). (The values here are chosen to show the arithmetic, not to report any single trial's exact figures.)

Discussion

Acceptance and Commitment Therapy occupies an unusual position in clinical psychology: it is at once a widely practiced treatment and an applied test of a basic theory of cognition. Its distinctive commitments — that the struggle against inner experience is more remediable than the experience itself, that language is the engine of much human suffering, and that a broadened, values-guided repertoire rather than symptom relief is the proper aim of therapy — follow from relational frame theory rather than being assembled pragmatically from what appears to work (Hayes, 2004). This theoretical ambition is the source of both its appeal and the sharpest criticism of it, since a therapy that claims a mechanistic basis invites the demand that the mechanism be demonstrated and not merely described.

On outcome, the fair summary is that ACT is an effective treatment for several conditions when set against doing nothing, and a treatment of contested advantage when set against established alternatives (Gloster et al., 2020). The tension in the literature is less about whether ACT helps than about how much of its benefit is specific to it, and the methodological critiques that pressed this question improved the field by raising the evidentiary bar rather than by dismissing the therapy (Öst, 2014). The process-level evidence, showing that isolated flexibility components move outcomes in the predicted direction, is the strongest support for the claim that ACT works through the mechanism it advertises (Levin et al., 2012).

The wider significance for cognitive psychology lies in the model's transdiagnostic logic. By locating pathology in a functional relationship to experience — avoidance and fusion — rather than in the content of particular symptoms, ACT reframes disorders as regional expressions of a common process, and offers psychological flexibility as a general-purpose target. Whether that reframing ultimately proves more useful than diagnosis-specific accounts is unsettled, but it has already shifted how a large part of the field thinks about what a therapy is trying to change (Kashdan & Rottenberg, 2010).

Current Directions

The most consequential recent development is the generalization of ACT into process-based therapy: an attempt to move clinical science away from named treatment packages and named disorders toward a direct focus on the biopsychosocial processes of change that any intervention must engage. On this view ACT is one historically important instance of a broader project, and its six processes are folded into a larger extended evolutionary account of how repertoires vary, are selected, and are retained across levels and dimensions of human functioning (Hayes et al., 2020). The reframing responds directly to the field's protocol proliferation by asking which processes a given case requires rather than which brand-name package to apply.

A second active line is methodological consolidation. Reviews of the meta-analytic literature have called for larger trials, stronger active comparators, and better mediation designs, and have begun to hold the ACT evidence base to those standards rather than to the weaker benchmarks of its early years (Gloster et al., 2020). The same period has seen sustained work on delivery format — group-based, brief, and internet-delivered protocols — that widens access while raising fresh questions about how much of the model's effect survives when its experiential exercises are compressed or automated (Coto-Lesmes et al., 2020).

Underlying both is continued attention to measurement and mechanism. Because the therapy is defined by process, the validity of its central instruments — above all the self-report indices of psychological inflexibility — remains an open and actively researched question, since a measure that cannot cleanly separate inflexibility from general distress limits any mediation claim built on it (Bond et al., 2011). The common thread across the current work is a turn from asking whether ACT as a package works toward specifying, measuring, and testing the processes through which it is supposed to work.

Common Misconceptions

ACT tries to reduce or eliminate negative thoughts and feelings.

It does the opposite. ACT aims to change a person's relationship to inner experience, building willingness to have unwanted thoughts and feelings while acting on values, rather than trying to lower their frequency or intensity (Hayes et al., 2006).

Acceptance means resignation or giving up.

Acceptance in ACT is active and in the service of action. It means making room for difficult experience so that behavior can move toward what matters, not passively tolerating a situation or abandoning efforts to change what can be changed (Hayes et al., 1999).

ACT is just mindfulness with a new label.

Mindfulness practices appear in ACT, but they serve two specific processes, present-moment contact and defusion, within a six-process model grounded in relational frame theory and organized around values and committed action (Hayes, 2004).

ACT is proven superior to cognitive behavioral therapy.

The evidence does not support that. ACT reliably beats inactive controls, but its advantage over established active treatments is small and contested, so claims of superiority outrun the data (Öst, 2014).

Glossary

Acceptance.

Active, willing contact with private experiences such as thoughts, feelings, and sensations without attempting to change their frequency or form, undertaken in the service of valued action.

Cognitive defusion.

The process of noticing thoughts as ongoing verbal events rather than as literal truths, weakening the behavioral grip a thought exerts without disputing its content.

Cognitive fusion.

The tendency to respond to a thought as though it were the event it refers to, so that verbal content dominates behavior; the inflexible opposite of defusion.

Committed action.

The building of larger patterns of effective behavior in the service of chosen values, sustained in the presence of the unwanted experiences that acceptance makes room for.

Contextual behavioral science.

The research tradition, rooted in functional contextualism and relational frame theory, from which ACT was developed and within which its processes are studied.

Experiential avoidance.

The attempt to alter the form, frequency, or situational sensitivity of unwanted private experiences even at behavioral cost; the process ACT places at the center of psychopathology.

Functional contextualism.

The philosophy of science underlying ACT, which analyzes behavior in terms of its function within a context and takes successful working, not literal truth, as its criterion.

Hexaflex.

The hexagonal diagram of the six core flexibility processes — acceptance, defusion, present-moment contact, self-as-context, values, and committed action — used to depict the model.

Present-moment contact.

Flexible, voluntary attention to what is happening in the here and now, as against attention captured by a conceptualized past or future.

Psychological flexibility.

Contacting the present moment as a conscious being and persisting in or changing behavior in the service of chosen values; the overarching goal of ACT.

Psychological inflexibility.

The rigid, avoidance-dominated pattern that results when the six processes sit at their inflexible poles; the target that ACT works to loosen.

Relational frame theory.

A behavior-analytic account of language and cognition holding that derived relations among events let arbitrary stimuli acquire one another's functions; the basic-science basis of ACT.

Self-as-context.

The experience of an observing, continuous self distinct from the thoughts, feelings, and roles it observes, providing a stable vantage from which difficult experience can be held.

Third-wave therapies.

Contextual cognitive behavioral approaches, including ACT, that target a person's relationship to thoughts and feelings rather than the content of cognition emphasized by earlier cognitive therapy.

Thought suppression.

The deliberate attempt to not think a particular thought, which experimentally produces an ironic rebound of the thought; a laboratory model for why experiential avoidance backfires.

Values.

Freely chosen, verbally constructed life directions that give ongoing behavior meaning and organize committed action, distinct from goals, which can be completed.

Key Researchers

Dermot Barnes-Holmes. Behavior analyst who co-developed relational frame theory, the account of derived relational responding that supplies ACT with its theory of how language forges the cognitive fusion the therapy targets. ORCID - Wikipedia - Wikidata

Frank W. Bond. Psychologist at Goldsmiths, University of London, who led development of the Acceptance and Action Questionnaire-II, the standard self-report index of psychological inflexibility, and extended ACT into occupational health. Google Scholar - Faculty page

Steven C. Hayes. Psychologist at the University of Nevada, Reno, who originated Acceptance and Commitment Therapy and relational frame theory and named the third wave of behavior therapy; the central architect of the psychological-flexibility model. ORCID - Wikipedia - Wikidata

Michael E. Levin. Psychologist at Utah State University whose component and meta-analytic studies isolated the active processes of the flexibility model and who has advanced web-delivered forms of ACT. Google Scholar - Faculty page

Kirk D. Strosahl. Co-founder of Acceptance and Commitment Therapy and co-author of its foundational text who adapted the model for primary-care and brief-intervention settings. ORCID - Wikidata

Kelly G. Wilson. Co-founder of Acceptance and Commitment Therapy and co-author of its foundational text whose work developed the values and committed-action components of the model. Google Scholar - Wikidata

Lars-Göran Öst. Clinical psychologist at Stockholm University whose methodologically rigorous meta-analysis of ACT trials tempered early efficacy claims and set a higher evidentiary bar for the therapy. ORCID - Wikidata

Frequently Asked Questions

What is Acceptance and Commitment Therapy?
It is a contextual cognitive behavioral therapy that treats suffering as arising from a rigid, avoidant relationship to inner experience, and that works to raise psychological flexibility, the capacity to accept unwanted thoughts and feelings while acting on chosen values (Hayes et al., 2006).

How is ACT different from cognitive behavioral therapy?
Standard cognitive therapy tries to change the content of dysfunctional thoughts, whereas ACT leaves the content alone and changes a person's relationship to thoughts, reducing the struggle to control them rather than correcting them (Hayes, 2004).

What are the six core processes of ACT?
They are acceptance, cognitive defusion, present-moment contact, self-as-context, values, and committed action, drawn together as the hexaflex and grouped into open, centered, and engaged response styles (Hayes et al., 2006).

What is psychological flexibility?
It is the capacity to contact the present moment as a conscious being and to persist in or change behavior in the service of chosen values, and it is the central goal that ACT aims to build (Kashdan & Rottenberg, 2010).

What is experiential avoidance?
It is the attempt to control, suppress, or escape unwanted private experiences even when doing so causes behavioral harm, and ACT treats it as a central process that brings short-term relief while narrowing a life over time (Hayes et al., 1996).

What is relational frame theory?
It is a behavior-analytic account of language holding that people derive untrained relations among events, which lets words acquire the functions of what they refer to and so lets language generate distress; it is the basic-science foundation of ACT (Hayes, Barnes-Holmes, & Roche, 2001).

Does ACT actually work?
Meta-analyses show ACT outperforms inactive control conditions across several disorders, but its advantage over established active treatments is small and contested, so its effectiveness is best stated relative to the comparison condition used (Gloster et al., 2020).

What conditions is ACT used for?
Because it targets a transdiagnostic process rather than a single disorder, ACT is applied across anxiety, depression, chronic pain, substance use, and other presentations, with the strongest review evidence for anxiety and depression (Twohig & Levin, 2017).

References