The big idea
B.19 and B.20 teach you to classify verbal responses by what controls them, and that is where the exam stops. In practice, Skinner’s (1957) analysis is also a teaching plan: if mands, tacts and intraverbals have different controlling variables, each has to be taught under its own conditions. Sundberg and Michael (2001) argued that this view improves autism treatment through better language assessment, early mand training, attention to establishing operations and a stronger intraverbal repertoire.
The “verbal behavior approach” is not a separate science. It is ABA that uses Skinner’s functional categories to choose and sequence language targets. This guide covers how the operants relate, how control moves between them, two ways of being verbal, and three concepts behind complex language: autoclitics, joint control and naming.
Breaking it down
Functional independence
Skinner proposed functional independence: a word learned as one operant does not automatically appear as another. Lamarre and Holland (1985) taught preschool children to mand or to tact with the same phrases (“on the left,” “on the right”) and found the two repertoires independent during acquisition. Later studies found more transfer: after mand training, all four preschoolers in Petursdottir, Carr and Michael (2005) also tacted, and three of four children with autism in Finn, Miguel and Ahearn (2012) showed immediate transfer between operants. The practical rule is to probe the untaught operant and teach it directly when it does not emerge.
Mand training
The mand is often taught early because it benefits the learner directly and can replace problem behavior that serves the same function. Mands need a motivating operation, so trainers capture MOs as they occur or contrive them:
- Interrupted chain: remove an item the learner needs to finish a familiar routine. Hall and Sundberg (1987) found that mands appeared reliably only after direct mand training under conditioned establishing operations, even though the same words were already tacts.
- Missing item: Sundberg, Loeb, Hale and Eigenheer (2002) hid a wanted item to teach “Where?”, then contrived a need for a specific person to teach “Who?” These are mands for information, and the children with autism used them with untrained items and in the natural environment.
- Withholding: briefly delay a routine item or action, such as the next push on a swing.
Transferring control between operants
Most programs start with a response the learner already emits and move it to a new controlling variable (G.8):
- Echoic to tact: say “car” while showing a car, then fade the echoic prompt until the car alone evokes “car.”
- Tact to mand: tact a preferred item while it is visible, then require the response when the item is wanted but out of sight.
- Tact to intraverbal: ask “What says moo?” with a picture of a cow, then fade the picture.
When these steps stall, adding echoic training can help (Kodak & Clements, 2009).
Intraverbals and verbal conditional discriminations
Intraverbals often lag behind mands and tacts. In an 80-item intraverbal subtest given to 39 typically developing children and 71 children with autism, most errors came from compound verbal conditional discriminations: questions in which two parts of the stimulus must control the answer together (“What do you eat?” vs. “What do you drink?”). Both groups made the same kinds of errors at the same score levels, which suggests teaching simple intraverbals before multi-part questions (Sundberg & Sundberg, 2011).
Topography-based and selection-based verbal behavior
Michael (1985) distinguished two ways of being verbal. In topography-based verbal behavior, each word has its own response form. In selection-based verbal behavior, the response is the same each time, and what differs is the stimulus selected.
| Topography-based | Selection-based | |
|---|---|---|
| Examples | Speaking, signing, writing | Pointing to pictures, exchanging icons, touching a speech-generating device |
| Response form | Different for each word | The same each time; the stimulus selected differs |
| Requires | Shaping each form | A conditional discrimination and scanning |
Michael questioned heavy reliance on selection-based systems. A review of 28 studies comparing communication systems found clear, consistent differences rare and recommended choosing from each person’s skills, goals and preferences (Gevarter et al., 2013).
The autoclitic in depth
An autoclitic depends on the speaker’s own primary verbal behavior and modifies its effect on the listener. Skinner described several kinds:
- Descriptive: tells the listener about the strength or source of the response (“I think,” “I heard”).
- Qualifying: negates or asserts it (“not,” “definitely”).
- Quantifying: “all,” “some,” “a,” “the.”
- Relational: grammar and word order that relate responses to one another.
An autoclitic frame is a learned frame with an open slot (“I want ___,” “The ___ is on the ___”) that combines with new responses to produce untaught sentences. Owen and Rodriguez (2024) taught children with autism autoclitics that signal weak stimulus control (such as “like a ___”), with generalization to untaught stimuli.
Joint control and naming
Joint control (Lowenkron, 1998) occurs when a response being rehearsed under one stimulus is evoked at the same moment by a second, and that coincidence sets the occasion for the next response. Told “Get the spoon,” a learner repeats “spoon” while searching and picks one up when the sight of it evokes the same word. After learning to rehearse instructions aloud, three children with autism followed new two-step instructions on the first trial (Vosters & Luczynski, 2020).
Naming (Horne & Lowe, 1996) joins listener and speaker behavior so that learning one produces the other (B.21, G.19, J.2). Petursdottir and Carr (2011) found little evidence for the common rule of finishing listener (receptive) training before speaker (expressive) training, and some evidence that the reverse order may be more efficient.
Seeing it in practice
| Situation | Verbal behavior reading | What to do |
|---|---|---|
| Tomas tacts 50 items but never asks for any | Functional independence | Contrive MOs (interrupted chains, withheld items) and transfer control from tact to mand |
| Jae answers both “What do you eat?” and “What do you drink?” with “cookie” | Only part of the question controls the answer | Teach the verbal conditional discrimination with contrasting question pairs |
| Ana’s only system is a 40-picture board, and she scans slowly | Selection-based behavior limited by scanning | Assess sign or a speech-generating device with her team, and ask her preference |
| A learner told “Get the cup” forgets the item before reaching the shelf | The instruction loses control during the search | Teach rehearsal of the instruction while searching (joint control) |
Common mix-ups
- Functional independence is a starting point, not a law. Probe untaught operants; sometimes they emerge.
- Selection-based does not mean nonverbal. Pointing to a picture can be a mand or a tact, classified by function.
- Autoclitics are not just grammar. They include any response that modifies how the listener reacts to the primary response.
- Receptive-first is a custom, not a finding. The evidence does not require listener training to come first.
Exam tip: The exam stops at classifying operants (B.19), multiple control (B.20) and emergent relations (B.21). A question about fading an echoic prompt until the object alone evokes the word is testing transfer of stimulus control (G.8).
Quick recap
- Each operant has its own controlling variables, so probe and teach each one.
- Teach mands under captured or contrived MOs, including mands for information.
- Transfer control from echoic to tact, tact to mand and tact to intraverbal.
- Intraverbal errors often come from compound verbal conditional discriminations.
- Choose topography-based or selection-based systems by the person’s skills and preferences.
- Autoclitics, joint control and naming account for much of complex, generative language.
Further reading
- Gevarter, C., O’Reilly, M. F., Rojeski, L., Sammarco, N., Lang, R., Lancioni, G. E., & Sigafoos, J. (2013). Comparing communication systems for individuals with developmental disabilities: A review of single-case research studies. Research in Developmental Disabilities, 34(12), 4415-4432.
- Lamarre, J., & Holland, J. G. (1985). The functional independence of mands and tacts. Journal of the Experimental Analysis of Behavior, 43(1), 5-19.
- Lowenkron, B. (1998). Some logical functions of joint control. Journal of the Experimental Analysis of Behavior, 69(3), 327-354.
- Michael, J. (1985). Two kinds of verbal behavior plus a possible third. The Analysis of Verbal Behavior, 3, 1-4.
- Petursdottir, A. I., & Carr, J. E. (2011). A review of recommendations for sequencing receptive and expressive language instruction. Journal of Applied Behavior Analysis, 44(4), 859-876.
- Sundberg, M. L., & Michael, J. (2001). The benefits of Skinner’s analysis of verbal behavior for children with autism. Behavior Modification, 25(5), 698-724.
- Sundberg, M. L., & Sundberg, C. A. (2011). Intraverbal behavior and verbal conditional discriminations in typically developing children and children with autism. The Analysis of Verbal Behavior, 27(1), 23-43.
Key terms
- Functional independence
- The finding that a response form learned as one verbal operant does not automatically occur as another, because each operant has its own controlling variables. It is tested by probing an untaught operant after another has been trained.
- Mand for information
- A mand, usually a wh- question, evoked by a motivating operation for some piece of information and reinforced by receiving it.
- Transfer of stimulus control
- The shift of control over a response from a prompt to the natural SD, achieved by gradually removing the prompt. Transfer is complete when the response occurs reliably to the SD alone.
- Verbal conditional discrimination
- A discrimination in which two or more parts of a verbal stimulus must control the response together, so that changing one word changes the correct answer.
- Selection-based verbal behavior
- Verbal behavior in which the response form is the same each time (pointing, touching or exchanging) and what differs is the stimulus selected. It requires a conditional discrimination and a scanning repertoire (Michael, 1985).
- Autoclitic
- Secondary verbal behavior that depends on the speaker's own primary verbal behavior and modifies how the listener reacts to it. Skinner described descriptive, qualifying, quantifying and relational autoclitics.
- Joint control
- Control of one response topography by two stimuli at once: a response being rehearsed under one stimulus is evoked at the same moment by a second, and that coincidence sets the occasion for the next response (Lowenkron, 1998).
- Naming
- A two-way relation between speaker and listener behavior. After learning to tact an object, a person can also select it as a listener when hearing its name, and the reverse, without separate teaching.
Check yourself
4 questions on J.13. Choose an answer to see whether you're right and why.
Show the answer
Answer: A. A response form learned under the control of a nonverbal stimulus did not occur under the control of an MO, which is what Lamarre and Holland (1985) called functional independence. The fix is to teach the mand directly, for example by transferring control from the tact. Divergent control means one variable evoking many responses, and nothing here shows the tact weakening.
Show the answer
Answer: B. A missing item that Teo wants makes its location valuable, which is the MO for "Where?" (Sundberg et al., 2002). Asking the question sets up an intraverbal, prompting with no missing item teaches an echoic, and praising pointing reinforces a listener response.
Show the answer
Answer: A. In selection-based verbal behavior the response (pointing) has the same form every time, and what differs is the stimulus selected, which requires a conditional discrimination and scanning (Michael, 1985). Signing, speaking and writing are topography-based: each word has its own form.
Show the answer
Answer: B. Joint control occurs when a response being rehearsed under one stimulus (the spoken instruction) is evoked at the same moment by another (the picture), and that coincidence sets the occasion for selecting (Lowenkron, 1998). There is no MO for the word, no question asked, and no prompt being faded.