Contribution
A compact statement of Relational Frame Theory by two of its founders, aimed at clinicians. It adds an argument for why the theory changes behaviour therapy, and a first RFT reading of three ACT processes. It reports no new data. Its contribution is the mapping from RFT’s basic concepts to clinical phenomena (trauma talk, anticipatory fear, metaphor, rules, the self), with worked examples for each.
Key insight
Animals readily learn that a neutral event predicts an important one, but not the reverse. Humans learn both directions, and then derive relations they were never taught. Once a word stands in a learned, bidirectional, derivable relation to an event, the word can carry the event’s psychological functions, and those functions can travel along relations (sameness, comparison, opposition) to events never encountered. So most human suffering is not only conditioned but verbal. Treatment has to work on the relations and the contexts that control them, not only on direct contingencies.
Assumptions
- Relating is operant. Derived relational responding is “a type of generalized operant behavior that has been brought under the control of contextual cues (e.g., the word ‘is’) through a process of differential reinforcement” (p. 361). The paper argues for this from training studies it cites; it is the theory’s central empirical claim.
- Continuity has limits. The “continuity assumption”, that principles found with rats and pigeons apply to humans, “has served its time” (p. 359). A human-specific research agenda is needed.
- Functional, not syndromal, classification. Behaviour analysts prefer dimensions such as emotional avoidance that “directly guide assessment and treatment” to DSM syndromes (p. 356, citing LIT-551).
Key results
These are conceptual claims illustrated by examples and supported by cited studies, not results the paper establishes.
- Bidirectionality (pp. 359–360). A dog learns “walkies” → exercise but not exercise → “walkies”; a child told after play that her toy is “Nimbo” is excited later by “Nimbo”. So a human’s report of a shock can acquire the shock’s aversive functions, and a pigeon’s report cannot.
- Networks and derived relations (pp. 360–362). Taught “Nimbo” = toy and “Daxy” = “Nimbo”, a child relates “Daxy” to the toy (equivalence). Multiple-exemplar training establishes the generalized operant. The same holds for more/less, different, opposite: a one-euro coin is “worth more” than a physically larger fifty-cent coin.
- Frames and their properties (pp. 362–363). Frames of coordination, comparison, distinction, opposition and hierarchy. Each has mutual entailment (A same as B → B same as A), combinatorial entailment (A > B, B > C → A > C and C < A; A > B and A > C leave B–C unspecified) and transformation of function, all under contextual control (“Think of” versus “Think of the smell of”).
- Clinical transformations (pp. 363–364). A child conditioned to fear “doctor”, told dentists are like doctors, fears the dentist without an aversive dental history. An agoraphobic man ranks Macy’s above the store where he panicked, through comparative frames (“larger than”, “busier than”).
- Relating relations (pp. 364–365). “Struggling with anxiety is like struggling in quicksand” relates two relational networks. The same process can make a person who feels “trapped” in a relationship sensitive to enclosed spaces. Skinner’s extended tact “fails to specify” how this develops.
- Rules (pp. 365–367). A rule is a relational network; following it is regulated by coordination between that network and the network describing what is happening. Rules understood may still not be followed, for four reasons: missing repertoire, speaker credibility, the speaker’s control of reinforcement, and the rule’s plausibility.
- The self (pp. 367–369). Verbal self-knowledge makes a passed-over promotion something one “may literally fail everyday”. Three selves: the conceptualized self (content of verbal relations), the knowing self (ongoing process) and the transcendent self (context). Decisions taken from self as context contact both content and process without either having “absolute control”.
- ACT (pp. 369–371). Acceptance: a rule predicting an aversive event (going to the mall → panic) that then coheres with experience; relational coherence is reinforced, so acceptance strategies persist. Defusion: exemplar training that shifts thoughts from I HERE NOW to I THERE THEN. Values: the top of a hierarchical relational network.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Human stimulus relations are bidirectional in a way nonhuman conditioning is not | moderate | cited studies (Leader, Barnes & Smeets 1996; Hall 1984); examples |
| C2 | Derived relational responding is a generalized operant learned from multiple exemplars | moderate | cited training studies (Barnes-Holmes et al. 2001a, b); argued, not shown here |
| C3 | Psychological functions transform through derived relations, which explains fear without direct conditioning | moderate | cited transformation studies (Dymond & Barnes 1995; Roche & Barnes 1997); clinical examples are illustrations |
| C4 | Metaphor is relating relations, and can both cause and treat distress | weak | conceptual; Stewart & Barnes-Holmes cited |
| C5 | Rule following is coordination between relational networks | weak–moderate | conceptual analysis of a worked example |
| C6 | Acceptance, defusion and values can be given RFT interpretations | weak | explicitly “no more than a heuristic” (p. 369) |
| C7 | Direct-contingency behaviour analysis is inadequate to human psychopathology | moderate as argument | the paper’s conclusion, resting on C1–C3 |
Concepts
- derived relational responding / arbitrarily applicable relational responding: relating events under the control of contextual cues set by the verbal community, not by the events’ formal properties (p. 362).
- relational frame: a pattern of derived relational responding; coordination, comparison, distinction, opposition, hierarchy (p. 362).
- mutual entailment, combinatorial entailment, transformation of function: the three defining properties of a frame (p. 363).
- relating relations: relating one relational network to another; the RFT account of analogy and metaphor (p. 364).
- rule-governed behavior: “a frame of coordination between two relational networks serves as a source of control over behavior” (p. 366).
- conceptualized, knowing and transcendent self: self as content, process and context of verbal relations (p. 368).
- I HERE NOW / I THERE THEN: the deictic positions defusion moves a thought between (p. 370).
Connections
- The RFT book (LIT-549), unread. Every concept here is cited to
its chapters. This paper is a précis of it with a clinical slant, and the
extendsrelation on the LIT says so. - Experiential avoidance (LIT-551), unread. Cited for the functional-dimensional alternative to syndromal diagnosis. This paper gives the mechanism the 1996 paper’s dimension needs: if words carry the functions of what they describe, private events can be as aversive as the events themselves, and avoiding them becomes a general strategy.
- The ACT model (LIT-555), read. The 2006 review condenses the same account into four implications and a model of “psychological inflexibility”. This paper’s acceptance, defusion and values are three of that review’s six processes. Its self as context is a fourth.
- Bandura (LIT-545), unread. Cited as one of the cognitive– behavioural integrations that kept behaviour therapy going after Chomsky without changing its basic science (p. 358).
- Emotion regulation (LIT-554). Gross’s process model treats regulation as acting on an emotion’s generation. This paper locates much of the trouble in the verbal relations that give private events their functions, which is a different place to intervene. This paper does not cite Gross.
Bearing on the record
- The
self-and-personhoodcluster. The three selves are a behaviour-analytic, functional account of the self: the self as what verbal relations are about, as their ongoing activity, and as the perspective (“I HERE NOW”) from which they are made. That is a kind of account the record’s other filings under that topic, which are philosophical, do not yet hold; this reading did not compare them. - A candidate THEORY, with the 2006 review: “verbal relations transfer aversive functions to private events, which makes experiential avoidance a general learned strategy”. This paper gives the mechanism, and the experimental evidence it cites would have to be read for status Active. See the batch report.
- No ML instruction. Nothing here belongs in the anthology.
- The three selves are compared with the record’s self holdings in THEORY-044. The clinical claim is THEORY-051.
Limitations
- No data of its own. Every empirical claim is a citation. The clinical cases are constructed examples.
- Part 3 is labelled heuristic by the authors (p. 369). The acceptance account, which rests on reinforcement of relational coherence, is not tested here.
- A partisan statement. The authors are the theory’s originators, and the paper does not engage the critics; their reply (2003) does.
- The DSM contrast is brief. The paper dismisses syndromal classification in a paragraph, citing its own tradition.
Open questions
- Is derived relational responding learned, as RFT claims, or does multiple-exemplar training only reveal a capacity it does not create? The paper asserts the first and cites training studies. The question is the centre of the commentaries the editors answered in 2003.
- Does relational coherence explain why acceptance-based instructions work, or would a simpler expectancy account do as well?
- Can the transformations of function the paper describes be measured in clinical populations, not only in the laboratory?
Corrections
- No DOI. IJPPT assigned none in 2004; the record carries the publisher’s URL.
- Pagination. The Maynooth deposit is a proof whose first page reads “pp. XXXXXX” and whose header page numbers start at 162; the published pages are 355–375. Page numbers here are the published ones.