Relational Frame Theory: How Language Builds Meaning
Relational frame theory explains how people relate things through language. Learn derived relations, entailment, and what RFT means for ABA practice.
Key takeaway
Relational frame theory (RFT) is a behavior-analytic account of human language and thinking. It says people learn to relate almost anything to anything else, and those relations change how we act and feel.

Playing the Game of Life with ACT: Forming Networks in the Real World
Free learning path on this topic
Verbal Behavior: Beyond WH-Questions
Dana Pettus's three-part series on building generative verbal behavior. Why WH-question programs stall, how to establish speaker-listener repertoires, and how to scaffold advanced verbal skills like recall and simple inference. Three hours, three CEUs.
Experts who cover this
11 recordings on openceu cover relational frame theory.
- Featured talk1 Learning CEU
Playing the Game of Life with ACT: Forming Networks in the Real World
Kaitlin WilderWildside Behavior LLC
Central framework of the session, explained as how learning a few relations lets the brain automatically derive a whole network of connected meanings that…
- 1 Learning CEU
Playing the Game of Life with ACT: Using RFT to Interrupt Networks in the Real World
Kaitlin WilderWildside Behavior LLC
Core framework of the talk: a behavior-analytic account of how humans automatically derive relational networks from experience without direct instruction…
- 1 Learning CEU
Practical RFT
Stu LawUpskill ABA
The central topic of the talk: Stu Law presents RFT as the account of why listeners listen (not just why speakers speak), arguing it explains language's power…
- 1.5 Learning CEUs
Language Changes the Game (Part 2)
Stu LawUpskill ABA
RFT is presented as the account of arbitrarily applicable relational responding (AARR); relational frames are described as operant behavior rather than stored…
- 1 Supervision CEU
Prosocial in the Workplace
Jason StaufferThe Behaviorist Bookclub
Explains relational frame theory as the basic science behind derived relational responding, illustrated with a spontaneous banana/power example and a…
- 1 Learning CEU
Real World” BCBAs, the “Behavior Crisis”, and Pathological Demand Avoidance
Kaitlin WilderThe Behaviorist Bookclub
Wilder cites relational frame theory as the basis of ACT, explaining it accounts for why people's verbal behavior and derived meaning often diverge from what…
- 1.5 Learning CEUs
The Four Function Paradigm (Part 1)
Stu LawUpskill ABA
Mutual entailment is explained as the ability to derive an inverse relation (e.g., if 'aunt is older than uncle' then 'uncle is younger than aunt'), and is…
- 1 Learning CEU
Multiple Exemplar Instruction (MEI): Training for Generalization
Dr. Jessica OsosCopper Consulting Group
Presented as a complementary framework to verbal behavior for explaining why MEI produces generative outcomes, with verbal behavior defining the function of…
- 1 Supervision CEU
Dunder Mifflin’s Guide to Training and Onboarding: Lessons from The Office
Mellanie PageThe ABA Collective, LLC.
Page uses RFT informally to explain how to teach RBTs to creatively generate new reinforcers by branching out from one known preference (e.g., an animal…
- 1 Supervision CEU
Beyond Competence: Using ACT to Cultivate Psychological Flexibility in Behavior Analytic Supervision
Falon RobersonRooted Consultation, LLC
Relational frame theory is cited as one of the behavioral foundations of ACT, explaining how language and learning history change the function of events for…
- 2.5 Learning CEUs
5 Days of Manding Mastery
Matt HarringtonThe Behaviorist Bookclub
Invoked to explain why concurrent mand-tact training works: verbal relations form networks, so learning to tact, mand, and use intraverbals together in the…
On this page · 19 sections▾
Relational frame theory (RFT) is a behavior-analytic account of human language and thinking. It says people learn to relate almost anything to anything else, and those relations change how we act and feel. In short, RFT explains how words get their power over behavior.
RFT grew out of Skinner's work on verbal behavior. It adds a focus on derived relations, which are links a person forms without being directly taught. For BCBAs and RBTs, it helps explain generative learning, rule following, and why talking to clients can change what they do. The experts below each come at it from a different angle.
RFT extends Skinner's account of verbal behavior#
Dr. Tom Szabo starts with history. He frames RFT as a next step after Skinner, not a rejection of him. Skinner defined verbal behavior as behavior reinforced through a listener who was trained to provide that reinforcement. Szabo points out that "speaker" and "listener" sat in quotation marks in Skinner's work. That detail matters, because Skinner's definition left room for odd cases, like pigeons trained in symbolic tasks.
What is relational frame theory? It's an approach that extends Skinner's way of conceptualizing verbal behavior.From the talk — Dr. Tom Szabo
Szabo then widens the frame. He calls RFT an account of "human languaging," meaning both language and cognition together. In his view, RFT lets us talk about thoughts and feelings as repertoires of behavior. A repertoire is simply the set of things a person can do. Once private events are framed as behavior, they can meet the standards of applied behavior analysis. That is the core promise Szabo sees. RFT gives behavior analysts a way to study thinking without leaving the science behind.
Why listeners listen, not just why speakers speak#
Stu Law makes a sharper claim about what RFT adds. Skinner's account focused mostly on the speaker. Law argues that RFT explains the listener side too. Why do people follow rules? Why does hearing a sentence change what someone does next? Those are listener questions.
rft adds why do listeners listen why do we follow rules what impact do does language have on what we doFrom the talk — Stu Law, Ph.D.
Law gives a pointed example. Skinner once said the link between an experimenter and a rat formed "a small but genuine verbal community." Law says that example "misses the point." Nothing about the rat's behavior changed. The pellets were just delivered by hand. For Law, the real question is what makes something verbal at all.
He also ties RFT to daily practice. Law says RFT shows "the limits to direct acting contingencies." Direct acting contingencies are consequences that shape behavior by contacting it directly. Language can bypass them. His example is preference assessments. You would never run a pairwise preference assessment with a teenage boy. You would ask him what he likes and probe from there. Language already does much of that work.
Arbitrarily applicable relational responding#
In his second talk, Law states the core claim of RFT in one line. People can do something called arbitrarily applicable relational responding, often shortened to AARR. That phrase sounds heavy, so it helps to break it apart.
relational frame theory has a very fairly straightforward claim rft argues that people human beings can do something called arbitrarily applicable relational responding sometimes referred to as r-ingFrom the talk — Stu Law
"Relational responding" means responding to how two things relate. "Arbitrarily applicable" means the relation does not depend on physical features. Law notes that animals like parrots can pick which object is bigger or darker. Those are physical relations. Humans go further. We can learn that a small coin is worth more than a big one, purely because of social training.
Law treats this as an ordinary skill. He compares it to an operant for opening doors, or to mands and tacts. A mand is a request, and a tact is a label. Generalized imitation is another comparison he uses. AARR is just one more part of a person's repertoire. Law says relational frames have five features. These include mutual and combinatorial entailment, control by a relational context, and transformation of stimulus function. The next sections unpack each one.
Relata, relational cues, and functional cues#
Szabo offers a clean way to map any relation. He uses a client statement: "I always make mistakes when I'm talking in a group." The two things being related are "I" and "making mistakes." RFT calls these relata, which is a fancy word for the things being related.
The word "always" tells you how they relate. It is a time relation stretched past every here and now. Szabo calls this the relational cue, or C-rel. It signals the type of link.
Then comes the context. "When I'm talking in a group" sets the situation where the relation holds. Szabo calls this a functional cue.
It may be the case that things get related in a particular way in one context, but not in another. So functional cues then are very important.From the talk — Dr. Tom Szabo
Szabo shows why this matters clinically. If the person loses the context, the thought becomes "I always make mistakes." Then they may stop doing anything at all. But if they notice it is only true in groups, not with a best friend or while walking the dog, that's different. Then they have room to act. They can practice group talk on a dog walk. Maybe they will make one or two mistakes instead of always failing.
Mutual and combinatorial entailment#
Entailment is the "free learning" at the heart of RFT. Mutual entailment means deriving the reverse of a taught relation. Law gives a family example from his talk on functional assessment. If you learn that your aunt is older than your uncle, you can derive that your uncle is younger than your aunt. That is the inverse of what was taught.
Notice the reverse is not identical. "Older than" flips to "younger than." That is why RFT goes past simple matching.
Combinatorial entailment means combining two taught relations to get a new one. Jason Stauffer walks through the classic case. Teach that A equals B, and the learner also knows B equals A. Teach A equals C, and the learner also derives C equals A. Then C equals B and B equals C come along too. As Stauffer puts it, "I teach you two. You get four for free." Law sums up the same idea in his own words: "humans are better at arbitrary relations we get stuff for free."
Stauffer adds that this speed may be uniquely human. He says we are the only species that can do this kind of thing so quickly.
Families of relational frames#
Equivalence, treating two things as "the same," is only one kind of relation. Stauffer stresses that language "isn't just about equivalence." He lists families of frames such as coordination, which links things as the same, and distinction, which marks things as different. He says the list is not complete. The number of frames is limited only by what the social verbal community trains.
Stauffer pauses on the deictic frame, since few people know the term. Deictic framing is about perspective. He calls it "kind of like me, you flips." What do I see? What do you see? Szabo also lists deixis among the frames and calls it "a fancy way of saying self."
Szabo's full list is wide. He names similarity, distinction, opposition (treating two things as opposites), equivalence, time, space, conditionality, and deixis. He argues that equivalence researchers did not anticipate how these frames interact.
Through relational frames, almost anything can be verbally related to almost anything else.From the talk — Jason Stauffer
Stauffer proved the point live. He asked the audience for two random words and got "banana" and "power." Then he asked how they relate. His answer: countries that grow bananas gain power because they have a good to trade. "Human beings are crazy at this," he said. We can take anything and link it to anything else.
Transformation of stimulus function#
Relating things would not matter much if it did not change behavior. Transformation of stimulus function means a stimulus takes on new effects through its relations. Nobody has to pair it directly with anything.
Stauffer describes a classic demonstration. Imagine two new squiggles. You learn that the S-shaped one is greater than the W-shaped one. Then the smaller symbol is paired with a mild shock. Soon your skin reacts to that symbol even with no shock. Here is the key part. Your skin now reacts to the greater symbol as if it means a bigger shock. That symbol was never paired with anything.
RFT is considered to be the mechanism that underlies this phenomenon.From the talk — Jason Stauffer
Law describes the same process in plainer terms. Language alters how we feel. It functions as reinforcers. It functions as antecedents. It "does all kinds of crazy stuff to our behavior."
Falon Roberson, writing about ACT in supervision, frames the RFT question in the same spirit. She says RFT asks how language and learning history are really changing the function of events. For a supervisor, that means a supervisee's thought like "I'm incompetent" has effects worth analyzing.
RFT and stimulus equivalence#
Many BCBAs first meet derived relations through stimulus equivalence. So a common question is whether RFT is the same thing. Szabo takes this on directly. He says the two are certainly related and emerged around the same time. But RFT is more than a description of how we relate things as "the same." It is a full account of language and cognition.
RFT is an account of human languaging. That is, of language and cognition.From the talk — Dr. Tom Szabo
Dr. Jessica Osos gives a practical way to hold both ideas. She treats equivalence as the entry point into RFT.
Stimulus equivalence is the most basic form of RFT, right?From the talk — Dr. Jessica Osos
Her example is simple. If you teach A equals B and B equals C, the learner gets A to C without teaching. Osos says this is "generativity instead of generalization." Generalization means a skill shows up in new settings. Generativity means brand new responses emerge that were never trained. That difference shapes how clinicians design programs.
Where experts differ: frames as behavior or as networks#
The speakers largely agree on what RFT explains. They differ more in how they talk about it. The sharpest contrast is about the word "network."
Law is firm that relational frames are behavior, not mental storage. He worries about words like "relational networks." Such terms can make frames sound like schema theory, where knowledge sits inside a person. For Law, frames live in a person's interaction with the environment.
relational frames are operant behavior. They're not stored in my brain any more than a kick or a thumbs up is stored in my hands.From the talk — Stu Law
Stauffer leans into network language more freely. He describes "millions of relational frames" built into "hundreds of thousands of relational networks." He borrows a biology idea here. Genes shape how a body presents, and our symbolic systems work like a "symbotype." That symbotype then shows up in our statements, beliefs, and values. Matt Harrington also speaks of "networks of verbal relations" when he explains concurrent mand and tact training.
These positions are not fully at odds. Stauffer still treats relational responding as learned. But the framing differs in a way that matters for teaching. Law's view keeps attention on contingencies you can arrange. Network and symbotype language is vivid and easy to grasp. Law's warning is that it can drift toward treating frames as things inside us. Clinicians who train staff on RFT should choose their words with that risk in mind.
What derived relations mean for functional analysis#
Law raises a striking clinical point in his talk on the four function paradigm. He cites research where experimental functional analyses were run with people who could show combinatorial mutual entailment. Once a person can give you the reverse of a relation, the odds of a clear FA result drop.
In Law's words, picture telling someone your aunt is older than your uncle. As soon as they can derive that your uncle is younger than your aunt, something changes. The chances of a standard FA yielding a result are "not good," Law says. He notes this is "a pretty simple task for a lot of people." An undifferentiated FA means no clear function shows up in the data.
Law ties this to the changing field. Behavior analysts had strong success when early intervention focused on ages three to five. Now BCBAs serve wider age ranges and more complex skills. Law mentions his own work with people in foster care. For verbally skilled clients, direct contingencies are only part of the story. Language and derived relations can shape behavior in ways a standard FA may not capture.
Multiple exemplar instruction and generative outcomes#
Osos uses RFT to explain why multiple exemplar instruction (MEI) works. MEI means teaching a skill across many varied examples and operants. She frames verbal behavior and RFT as partners, not rivals. Verbal behavior defines the function of each operant. RFT accounts for derived relational responding.
From an RFT view, Osos says MEI may strengthen the contextual cues that support relational responding. That makes it more likely that stimulus functions transform without direct training. She lists what effective MEI needs: consistent stimulus sets, rapid and randomized trials, and operant rotation. Consequences must also match each operant's function. She flags operant rotation as the piece most often missing. Without it, procedures slide back toward generalization rather than generativity.
Harrington reaches a similar conclusion from a manding study. When manding and tacting were taught at the same time, results were better than teaching one operant at a time. He links that finding to RFT. Learning to mand, tact, and use intraverbals in the same context fits what we know about networks of verbal relations. An intraverbal is a verbal response to someone else's words, like answering a question. Harrington adds that this matches his own practice. Teaching one operant at a time used to produce slow results for him.
Testing derived relations in skill programs#
Brian Middleton shares a mistake many clinicians may recognize. He was running PEAK-style programs for a learner stuck on coins. PEAK is a curriculum built around relational training. He ran discrete trial teaching (DTT) and then generalization. He never moved on to test the derived relations.
I'm doing the DTT and then the generalization. And I'm not moving into the transitivity and reflexivity sections. And I was making that mistake.From the talk — Brian Middleton
Transitivity is the combined link, like A to C after training A to B and B to C. Reflexivity is matching a thing to itself. Middleton fixed the problem by writing his own program. He trained a chain of relations, then tested whether the reverse and combined relations appeared.
The payoff went beyond one learner. Once he trained staff on the new design, they ran those programs better. They also started running PEAK itself more effectively. His advice is to use PEAK as a guide, not the final word. Build the relational connections on purpose, and test for them.
Using RFT to train RBTs, not just learners#
Mellanie Page applies RFT to staff training. She is open that she is not an RFT expert. Still, she sees value in how it explains learning that grows on its own.
Her problem is a common one. RBTs get stuck on one reinforcer. Johnny likes the iPad, so he gets the iPad. Then he satiates, and problem behavior can follow. Page found creativity hard to teach directly. RFT gave her a method.
Have the RBT then create branches of other activities or reinforcers that relate on some level to animals.From the talk — Mellanie Page
Her example starts with a child who loves an animal stuffy. The RBT asks what the child likes about it. Is it soft? Is it the role play? Is it that it's an animal? Each answer becomes a branch toward new reinforcers. Page's point is that creativity is not just an inborn trait. You can shape it by having RBTs practice expanding from one idea to many.
Learn it in one, derive it in two: networks in childhood#
Kaitlin Wilder teaches RFT through a two-part ACT series. In the first talk, her stated goal is to use RFT to understand how networks form in childhood. A network here just means a web of linked meanings. Those early webs shape how a person acts as an adult.
Wilder shares a rhyme she credits to Steve Hayes, one of the founders of RFT. She says she loves it as much as he does.
Learn it in one, derive it in two. Put it in networks. that change what you do.From the talk — Kaitlin Wilder
She unpacks the rhyme in plain words. When we learn one thing, the brain makes many more links on its own. No one has to teach them, and no direct experience is needed. It simply follows from how thinking works.
learning just a couple of relations allows you to instantly generate a web.From the talk — Kaitlin Wilder
Those webs then change the function of things around us. In other words, objects and events start to mean something new. That new meaning steers emotions, decisions, and behavior. Wilder says clients often show up "tangled up" in these webs. Her work aims to help them gain some logic about why they do what they do. She also uses journaling so people can spot patterns from their own history.
Noticing and interrupting networks#
In her second talk, Wilder moves from how networks form to how to interrupt them. Her objectives are to use RFT to notice the networks that drive observable behavior. Then she uses ACT to help people relate to those networks more flexibly.
She walks through the main frame families one at a time. These include coordination, opposition, distinction, comparison, hierarchy, space, and time. Comparison means ranking things as more or less. Hierarchy means nesting one thing inside a bigger category. For each one, she has learners journal about how it shows up in daily life. The focus is on three kinds of verbal products: assumptions, judgments, and rules. A rule is a verbal statement that tells us how to act, like "I must never be late."
Wilder also argues that this matters for practice. She says conventional ABA alone will not get clinicians far with these human problems. Behaviors driven by verbal networks need tools that reach the networks. This echoes Law's point about the limits of direct acting contingencies. It also fits Stauffer's shock example. A rule or judgment can change behavior without ever being paired with a real consequence.
How much RFT do you need to know?#
A practical question runs through several talks. Do clinicians need the full technical theory to use RFT well? The speakers lean the same way, though they phrase it differently.
Szabo raises the question head on. He notes that ACT has six processes and RFT has a lot of moving parts. He then asks how much of it a practitioner really needs. His answer points toward treating these processes as repertoires of behavior. If you can describe them that way, you can meet the standards of applied behavior analysis.
Wilder is blunt about the math side of RFT. She says formal RFT training can feel "very trigonometric." She loves math, but she does not believe that level of detail is necessary to apply RFT to real human life. Her program focuses on noticing frames in conversation instead.
Law takes a similar stance. He warns listeners that he speaks about RFT in ways that "might feel very loose." If you want strict technical definitions, he says, you may be disappointed. Still, he says RFT has let him predict and influence verbal behavior in his own practice. Page also admits she is not an RFT expert, yet she finds it useful. Together, these voices suggest a working grasp of the core ideas goes a long way.
RFT in everyday clinical life and ACT#
RFT is often called a foundation of acceptance and commitment therapy (ACT). Several speakers tie the two together. Roberson places RFT next to functional contextualism as one of ACT's behavioral roots. Functional contextualism is the idea that behavior is understood by what it does in context. Szabo's talk asks how much RFT a clinician needs to know to use ACT well.
Kaitlin Wilder speaks from the trenches. She describes the mind as "a meaning-making machine." You tell a caregiver something clear, and they leave and do something else. It can feel like they did not hear a word.
that's rft in real life right there that's exactly how the human mind and human behavior works and if we don't understand that then we're just going to be thinking we're a crazy person for our whole career.From the talk — Kaitlin Wilder
For Wilder, RFT reframes that frustration. The gap between what was said and what was done is not random. It reflects derived meaning shaped by each person's history. She argues many BCBA struggles come from this. Clinicians hand over a strategy, but the real problem is happening in the caregiver's verbal behavior. Seeing that clearly protects clinicians from burnout and blame. For a deeper look at how Law connects these ideas to practice, see Language Changes the Game (Part 2).
What the research says#
The research base on RFT is broad and still growing. A few recent papers stand out for practitioners.
Stricker and colleagues tested an online relational skills program called SMART with school children (Stricker et al., 2024). Students did 45-minute sessions one to three times a week for about ten weeks. Controlling for baseline scores, the intervention group scored significantly higher in math and reading than a matched control group. The control group also made modest gains. The paper notes that RFT places a range of derived relational operants at the core of human cognition.
Barnes-Holmes and colleagues revisit what a relational frame actually is (Barnes-Holmes et al., 2026). They stress that RFT explains frames with the same principles used for any operant. They cite Catania's point that training with many examples can produce generalized classes, sometimes called relational frames. That observation fits Osos's case for multiple exemplar instruction.
Regaço and colleagues compare stimulus equivalence, RFT, and naming theory (Regaço et al., 2025). They argue that a full behavioral account of language draws on all three.
A survey of practicing BCBAs shows a clear gap (Enoch & Nicholson, 2020). About 68% said they did not use RFT in their interventions. About half rated their own RFT understanding as low.
FAQ#
What is the difference between RFT and stimulus equivalence? Stimulus equivalence covers "same as" relations, like A equals B and B equals C. RFT includes equivalence but goes further. It covers opposition, comparison, time, space, and perspective relations. It also aims to explain language and cognition as a whole.
What does "derived relational responding" mean? It means responding to relations that were never directly taught. If you learn A is bigger than B, you can say B is smaller than A. Nobody trained that second answer. It was derived from the first.
How does RFT help BCBAs in daily practice? It helps clinicians design programs that test for untrained relations, not just trained ones. It explains why teaching operants together can speed learning. It also shows why language can change behavior and emotion, which matters for verbally skilled clients and caregivers.
Turn this topic into a CEU
You just studied this. Now get credit for it.
Watch Playing the Game of Life with ACT: Forming Networks in the Real World with Kaitlin Wilder and earn 1 free BCBA CEU. Detailed certificate, delivered the moment you finish.
Keep going with a free learning path
Verbal Behavior: Beyond WH-Questions
Dana Pettus's three-part series on building generative verbal behavior. Why WH-question programs stall, how to establish speaker-listener repertoires, and how to scaffold advanced verbal skills like recall and simple inference. Three hours, three CEUs.







