Equivalence, RFT, Naming, or Joint Control?: The Quantification of Private Events can Help Move us Forward Mark L. Sundberg Based on a paper by: Sundberg,

Equivalence, RFT, Naming, or Joint Control?: The Quantification of Private Events can

Help Move us Forward

Mark L. Sundberg

Based on a paper by:

Sundberg, C. T., Sundberg, M. L., & Michael, J. (under preparation)

Equivalence and Emerging Relations

• Sidman (1971)

• However, after 40 years of the accumulation of data consensus on the relevant sources of control responsible for the emergence of new behavioral relations remains elusive

• Dougher et al. (2014) open their Editorial introducing the recent special issue of The Journal of the Experimental Analysis of Behavior (JEAB) dedicated to stimulus-stimulus relations with, “One of the great challenges for a behavioral science is to provide an account of emergent stimulus-stimulus relations not explained by primary stimulus generalization”

Equivalence and Emerging Relations

• Four similar, but clearly distinct, conceptual frameworks that strive to provide an account of emergent stimulus-stimulus relations:

• Sidman’s (1994) “Equivalence Theory”

• Hayes, Barnes-Holmes, & Roach’s (2001) “Relational Frame Theory (RFT)”

• Horne and Lowe’s (1996) “Naming Theory”

• Lowenkron’s (1998) “Joint Control”

Equivalence and Emerging Relations

• There are many elements and issues that could be raised regarding the distinctions between these four alternative explanations

• But perhaps the most significant issue among them is the role that overt and covert mediating behaviors may play in evoking the target response (Horne & Lowe, 1996; Stromer, 1996)

• Common to use a MTS preparation with verbal participants

• Sidman/Hayes: direct effect of the manipulated contingencies• Common not to mention covert VB, not to attempt to measure it, or to dismiss it

• Horne & Lowe/Lowenkron: verbal mediation, rule governed• Focus on overt and covert mediating behavior as a source of control for emerging

relations

Explaining Emerging Behavioral Relations

• Horne & Lowe (1996) note that, “Verbally able subjects performing on matching-to-sample tasks often talk to themselves, sometimes overtly but most of the time covertly about the task and how they should perform on it” (p. 329)

• Stromer (1996) states “Horne and Lowe propose a ‘naming hypothesis’ of stimulus class formation that is arguably a verbal mediation account” (p. 250)

• Stewart, McElwee, & Ming (2013) reject naming and joint control due to their focus on mediation. These authors state:

• “the fact that naming and joint control both require an additional mediational process to explain derived stimulus relations in comparison to RFT can be seen as a weakness of the former” (p. 143)

• Sidman also rejected naming and mediating behavior as playing a role in equivalence (Sidman & Tailby, 1982; Sidman, et. al., 1986)

Explaining Emerging Behavioral Relations

• Discrepancy between those who easily demonstrate stimulus equivalence, and those who don’t (JEAB’s Jan., 2014 special issue)

• Existing verbal repertoires, verbal mediators, and a participant’s history seem relevant to outcomes

• The RFT and equivalence position reflects what Skinner (1974) identified as “methodological behaviorism”

• The RFT and equivalence position both call for significant changes in the basic concepts of behavior analysis

• For example, Hayes, Barnes-Holmes, & Roche (2001) state that it “is now time for behavior analysts to abandon many of the specific theoretical formulations of [Skinner] in the domain of complex human behavior … Many of the most prominent Skinnerian ideas about human complexity must be put aside or modified virtually beyond recognition” (p. xii)

Radical Behaviorism, Private Events and Multiple Control

• “With respect to each individual…a small part of the universe is private. We need not suppose that events which take place within an organism’s skin have special properties for that reason. A private event may be distinguished by its limited accessibility but not, so far as we know, by any special structure or nature” (Skinner, 1953, p. 257, see also Skinner, 1945, 1974)

• “the problem of privacy cannot be fully solved by instrumental invasion of the organism. No matter how clearly these internal events may be exposed in the laboratory, the fact remains that in the normal verbal episode they are quite private.” (Skinner, 1957, p. 130)

• Skinner suggests 4 ways a verbal community teaches its participants to identify and talk about private events (public accompaniment, collateral behavior, common properties, response reduction)

Radical Behaviorism, Private Events and Multiple Control

• Part of the complexity of quantifying private events is that in addition to their inaccessibility, they can take many forms (e.g., 1957, Ch. 19, Thinking)

• verbal (e.g., echoic, tact, intraverbal, mand, autoclitic)

• nonverbal (e.g., covert images of places or movement)

• motivating operations (e.g., aversive stimuli, sleep deprivation)

• discriminative stimuli (e.g., self as a listener)

• consequences (e.g., automatic reinforcement and punishment)

• respondent relations (e.g., fear, anxiety)

• These variables usually do interact with each other and experimenter manipulated variables in types of multiple control (Skinner, 1953, 1957; Michael, Palmer, & Sundberg, 2011)

Difficulties in Conducting Empirical Research on Complex Behavior

• “The analysis of complex cases” (Skinner, 1953, Chapter 14)

• Skinner’s first detailed presentation of “multiple control”

• “Two facts emerge from our survey of the basic functional relations: (1) the strength of a single response may be, and usually is, a function of more than one variable and (2) a single variable usually affects more than one response” (Skinner, 1957, p. 227)

• “when variables are combined in different ways…an important part of (our) task is to show how (the) variables interact” (1953, p. 205)

• In addition to multiple control, Donahoe & Palmer (2004) identify three general problems in the study of complex behavior

• 1) complex “behavior…varies from person to person” (p. 2)

• 2) “complex behavior…has its roots in the past, but knowledge of the past is incomplete” (p. 2)

• 3) “much ongoing activity is inaccessible to an observer” (p. 2)

Towards Quantifying Private Events

• In an effort to encourage the empirical investigation of private events, Palmer (2011, p. 203) suggested several indirect ways to quantify private events. Some of Palmer’s suggestions, along with others are…

• Collateral behavior

• Motivating operations: Interrupted chain procedure (Hall & Sundberg, 1987)

• EO and AO manipulations (Shillingsburg, Bowen, Valentino, & Pierce, 2013)

• Interfering tasks

• Pre-session verbal training (Wulfert, Dougher, & Greenway, 1991)

• Invent novel topographies• Making targeted textual stimuli unpronounceable (Mandell & Sheen, 1994)

• Comparisons that rhyme with the sample stimulus

• (Randell & Remington, 2006)

Towards Quantifying Private Events

• Blocking, conflicting, or confusing vocal stimuli

• Blocking covert self-rules (e.g., Taylor & O’Reilly, 1997)

• Talk aloud and exit interview procedures

• (Ericsson & Simon, 1984; Potter, Huber, & Michael, 1998)

• Overt mediators (e.g., echoic, vocal tacting, hand signs, arrows)

• (Lowenkron, 1984, 1988; Sundberg & Sundberg, 1990)

• The “silent dog” method

• (Hayes, et al, 1986; Arntzen & Halstadtro, 2009)

• Limited-hold contingencies

• (Holth & Arntzen, 2000)

Mediating Behaviors as Additional (Multiple) Sources of Control

• Ferster and Skinner (1957) defined mediating behavior as “Behavior occurring between two instances of a response being studied (or between some other event and such instance) which is used by the organism as a controlling stimulus in subsequent behavior” (p. 729)

• Sidman (1960) also discusses mediating behavior: “The subject will often adopt a posture in which his whole body or part of it maintains a constant position relative to the correct container…Such mediating behavior may enable the subject to select the correct container even after the lapse of a considerable amount of time” (Sidman, 1960, p. 375)

• Mediating behavior can be overt, covert, verbal, or nonverbal

• Overt nonverbal mediating behavior was extensively studied in the early behavioral literature (e.g., Blough, 1959; Eckerman, 1970; Hodos, Ross, & Brady, 1962; Laties, Weiss, Clark, & Reynolds, 1965; Shimp & Moffitt, 1977; Wilson & Keller, 1953)

Empirical Research on Mediating Behaviors

• For example, Blough (1959) showed that pigeons who emitted differential chains of stereotypic behavior during two different delay conditions demonstrated a higher frequency of correct responding versus pigeons who did not engage in such behaviors

• Blough concluded that “behavior during the delay interval seemed to determine the matching response” (p. 156)

• In addition, “incorrect responses typically followed the occurrences of the ‘wrong’ chain” (p. 157)

• Overt mediating behaviors “themselves provided discriminative stimuli for the matching response” (Blough, p. 157)

Precurrent Behaviors

• Skinner (1953) termed these mediating behaviors “precurrent behaviors” (p. 76)

• He made the point that precurrent behaviors can produce response products (e.g., visual, auditory, kinesthetic), can be public or private, verbal or nonverbal, and have multiple functions (e.g., SDs, MOs, Srs)

• For example, in his analysis of problem-solving behavior, Skinner (1969) states “The question ‘Who is that behind you?’ poses a problem….turning and looking are precurrent responses which generate a discriminative stimulus required in order to emit a particular name” (p. 142)

• For Skinner, the primary function of precurrent behavior was “mainly to make subsequent behavior more effective” (1968, p. 124)

Precurrent Behaviors

• Parsons and colleagues (Parsons & Ferraro, 1977; Polson & Parsons, 1994; Parsons, Taylor, & Joyce, 1981) extended the basic animal research on precurrent collateral behavior to humans

• For example, Parsons, et al (1981) showed that when 12 Kindergarten children in an MTS preparation were required to emit a specific collateral precurrent behavior they “rapidly learned the conditional discrimination” (p. 259)

• When precurrent behavior was not available “Subjects…either failed to acquire the discrimination or did so incrementally over a number of sessions” (p. 60)

• When precurrent behavior was prevented “these subjects showed marked decrements in matching” (p. 263)

Joint Control and Mediating Behaviors

• Lowenkron (1984, 1988, 1989) conducted a series of experiments where participants were taught overt responses capable of mediating generalized delayed matching (rotating an arrow, hand signs, use of a compass)

• It was demonstrated that generalized delayed matching was dependent on these mediating responses

• Lowenkon suggested a type of multiple control termed “joint control” is responsible

• Joint control occurs when two separate antecedents that evoke the same response form occur simultaneously (e.g., a self-echoic and a tact) and function as a discriminable event (a new SD) (Skinner, 1969)

Joint Control and Mediating Behaviors

• For example, when looking for a gate at an airport one might first look at the departure board and find the appropriate gate number (e.g., B85)

• The actual terminal and gate may be some distance away. A self-echoic prompt allows the traveler to retain the terminal and gate number over time and distractions

• As the traveler passes by gates he overtly or covertly tacts the numbers (“there is B81”) and continues on until his self-echoic matches the response form produced by the tact of the numbers

• When the response product of the self-echoic matches the response product of the tact, a new discriminable event occurs and it is this new SD that evokes selection behavior (going to that gate)

Joint Control

CovertSD

3/R1

Tact“Gate B85”

JointControl

SD4

Overt

R3

(Selection)

Overt

OvertSD

2/R2

(Scanning)

Covert

SD1/R1

Self-Echoic

“Gate B85”

Automatic SR+

“There it is!”

Covert Consequence

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

Establishes SD3

Gate B85

Go to that gate

Selection-Based Verbal Behavior and Topography-Based Verbal Behavior

Michael (1985) introduced the distinction between topography based verbal behavior (TBVB) and stimulus-selection based verbal behavior (SSBVB)

In TBVB each verbal operant involves a unique response topography (e.g., words, signs, spelling)

In SSBVB the response topography is the same (e.g., selection), but what differs is the stimulus selected (e.g., PECS, icon selection, MTS)

Sundberg and Sundberg (1990) examined distinction between TBVB and SSBVB with 4 low-verbal developmentally disabled adult participants

Various arbitrary MTS relations were established between auditory and visual stimuli with hand signs and with selection behavior

Selection-Based Verbal Behavior and Topography-Based Verbal Behavior

• The results showed that only the participants who were trained with the TBVB system demonstrated transitivity

• When TBVB conditions were changed to SSBVB conditions, transitivity did not occur

• When SSBVB conditions were changed to TBVB conditions, transitivity occurred

• These results were replicated and extended by Wraikat, Sundberg, and Michael (1991)

• Reflecting on these data, and his own, Lowenkron (1991) suggested the possibility that the stimulus selection-based response (as in MTS) “depended on mediation by a topography-based behavior” (p. 126)

Disrupting Mediation

• In Parsons, et al (1981) during the precurrent behavior prohibition condition, “subjects showed marked decrements in matching” (p. 263)

• Lowenkron (2006) showed that performance decreased when either the self-echoic or tact components of joint control were disrupted

• Gutierrez (2006) disrupted self-echoics by requiring the participants sing “Happy Birthday” between the sample and comparison stimuli and performance decreased

• DeGraaf & Schlinger (2013) also showed the negative effects of disrupting the self-echoic (reciting the alphabet, counting)

• In addition, they demonstrated that a joint control procedure was more efficient than a standard prompt and fade procedure in establishing conditional discriminations

Current Study• This study sought to identify the role of covert TB verbal mediation in

the acquisition of delayed arbitrary MTS with high verbal vs. low verbal participants

• A procedure that was designed to disrupt covert tacts and thus prevent the set-up of joint control was examined

• The same procedure was used with high and low-verbal participants• A standard arbitrary matching-to-sample procedure was designed to

establish various relations between • Nonsense words (e.g., “bibi,” “nadia,” “kayba,” “taida”)• Visual symbols (e.g., boxes, shapes, lines)• Odd objects (e.g., LEGO configurations)• Hand signs (e.g., a fist, wiggle fingers)

Method

• Two groups of participants:

• 4 college students

• 4 adults with developmental disabilities and limited verbal skills

• Tasks were balanced in difficulty, measured by error rate

• Repeated acquisition design was used

Method

• There were four conditions:

• Auditory-visual MTS with fixed position in the comparison array

• Auditory-visual MTS with random position in the comparison array

A Sample Comparison Array forCollege Students: Fixed

A Sample Comparison Array forCollege Students: Random

A Sample Comparison Array for Students with DD: Random

Verbal SD

Find “Bibi”

A Sample Comparison Array for Students with DD: Random

Verbal SD

Find “Bibi”

Method

• There were four conditions:

• Auditory-visual MTS with fixed position in the comparison array

• Auditory-visual MTS with random position in the comparison array

• Auditory-visual MTS with all comparison stimuli the same but in a fixed position

A Sample Comparison Array for College Students: Same Symbols, Position Fixed

Method

• There were four conditions:

• MTS with fixed position in the comparison array

• MTS with random position in the comparison array

• MTS with all comparison stimuli the same but in a fixed position

• Topography-based manual signs as tacts and intraverbals

• Followed by exit interviews and talk aloud procedures

The Same Condition

• All comparison stimuli the same, but in fixed positions

• Participants with DD performed as well in the “same condition” as they did in the “fixed condition”

The Same Condition

• All comparison stimuli the same, but in fixed positions

• Participants with DD performed as well in the “same condition” as they did in the “fixed condition”

• College students performed the worst in “same condition”

A Sample Comparison Array for College Students: Same Symbols, Position Fixed

Exit Interview and Talk Aloud

• All college students interviewed (3) reported “developing strategies” to find the correct comparison stimulus

• Strategies were very similar for all participants, also demonstrated in novel talk aloud post-experiment rounds

• Basic strategy was to alter the spoken nonsense word to something more familiar, and create a name for each of the symbols

• Then create a connection between the new word and the symbol

• Relation: “Bibi,” ---> square w/vertical and horizontal lines. One participant reported that “Bibi” was (intraverbally) converted to “baby” and the framed square was thought of (tacted) as a window, and the (intraverbal) phrase “baby in the window” was used to “connect” the two stimuli

Exit Interview and Talk Aloud

• Other examples:

• (“Biba,” ---> empty square) ---> Biba was converted to “Bebop” and the square was thought of as a dance floor

• (“Luba” ---> empty square) ---> Luba was converted to “Love life” then “my love life is empty”

• (“Quba” ---> square with a line though it) ---> Quba was converted to “Club,” then the symbol was thought of as a stick

Overt Covert

Tact“Window”

“Bibi”

Overt Covert

Echoic/Intraverbal

“Baby”

A Verbal Behavior Analysis of MTS with Verbal Participants: Early Trials

A Verbal Behavior Analysis of MTS with Verbal Participants: Later Trials

Covert

R2/SD3

Echoic/Intraverbal

“Baby in the Window”

Covert

R1/SD2

Echoic/Intraverbal

Sample SD

1

“Bibi”

Overt

“Baby”“Bibi”

A Verbal Behavior Analysis of MTS with Verbal Participants

Covert

R4

Tact“Window”

JointControl

SD5

Overt

R5

Selection

Overt

OvertR3

(Scanning)

Covert

R2/SD3

Echoic/Intraverbal

“Baby in the Window”

Points!

Consequence

Automatic SR+

“Aha!”

Consequence

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ SΔ

SΔ

SΔ

Establishes SD4

A Sample Comparison Array for College Students: Same Symbols, Position Fixed

Exit Interview and Talk Aloud

• All three participants reported that they could not use their covert strategies in the “same phase” to find the correct comparison stimulus and relied on repeated trials to learn the relations

What Happens in the “Same” Condition with Verbal Participants?

Covert

R2

Tact“Window”

JointControl

SD5

Overt

R2

Selection

Overt

OvertR1

(Scanning)

Covert

R2/SD3

Echoic/Intraverbal

“Baby in the Window”

Points!

Consequence

Automatic SR+

“Aha!”

Consequence

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ

SΔ SΔ

SΔ

Establishes SD2

Sample SD

1

“Bibi”

Overt

Conclusions

• Mediating covert verbal behavior (private events) can play a role in complex MTS tasks with verbal participants

• But, mediating covert verbal behavior played no role with the low-verbal participants

• McIlvane (2014) noted, “For some years now, results of the work with both nonhumans and nonverbal humans has begun to convince me that the matching-to-sample methodology that we all grew up with professionally will be shown to be unacceptably inefficient for answering certain types of questions” (p. 163)

• RFT, equivalence, naming, or joint control?

• The results from the current study support Lowenkron’s analysis of joint control, which can provide a more fine-grained analysis of emerging relations than Horne & Lowe’s naming relation

• The removal of the tact component prevents the set-up of joint control

Conclusions

• Results replicate and extend previous studies on disrupting joint control (DeGraaf & Schlinger, 2013; Gutierrez, 2006; Lowenkron, 2006)

• The call by RFT supporters to abandoned Skinner’s analysis of complex behavior and verbal behavior seems unwarranted

• Sidman’s suggestion that equivalence be considered a basic behavioral process also seems unwarranted

• Emerging relations can be explained with existing behavioral principles and concepts, thus, no need for any new principles (equivalence), higher order concepts (naming), or reformulation of behavior analysis as we know it (RFT)