Procedures for transfer of stimulus control are well suited to teaching tacts (Barbera

& Kubina, 2005) when echoic and mand repertoires are previously established and when
shaping and errorless learning procedures are incorporated (Partington et al., 1994).
Specific procedures are described below for structured work situations and play-based
interventions such as the Natural Language Paradigm (NLP; Koegel, ODell, & Koegel,
1987).

Transfer of Stimulus Control

REHFELDT
BARNES-HOLMES

Procedures for transfer of stimulus control establish responding under some version
of multiple control with either echoic prompts (for example, G.I. Joe), EO control (in
other words, manding), or both, and the presence of the nonverbal stimulus (in other
words, tacting). Targeting multiple controlled responses rather than pure operants may
result in higher compliance and greater enjoyment for the learner without sacrificing speed
of acquisition or overall strength of the learned repertoire (Braam & Sundberg, 1991).
Once responding is established, the extraneous controlling variables are faded until only
the nonverbal stimulus occasions responding. The clinician should initiate training with
the fewest possible extraneous controlling variables, because fading each variable requires
additional time and effort. Thus, a child with a strong echoic repertoire and reasonable
levels of compliance might only require echoic prompts (for example, Say boat in the
presence of a toy boat or a picture of a boat) without the need for manipulation of EOs
to occasion responding. Children with more limited echoic repertoires or who consistently emit problem behavior in response to demand situations might benefit more from
inclusion of EO control in establishing initial responding and might also perform more
consistently in a play context.
To use transfer of stimulus control procedures to teach tacts in a structured teaching
environment, first identify a robust verbal operant repertoire (for example, mand, echoic,
or intraverbal) and base your prompting strategies on that operant. Next, have the child
sit at a table and present the nonverbal discriminative stimulus (in other words, show the
item to the child). If you are using errorless prompting strategies, immediately provide the

Derived Relational Responding

Procedures for Teaching Basic Tacts

Psychology

relevant prompt (for example, if the verbal operant is echoic, the prompt might be red).
Respond to accurate responses with immediate praise and, if needed, either unrelated
reinforcers or theBreakthrough
item (in other words,
mand Behavior
transfer). Note
that as the
tact repertoire
Applied
Analytic
Techniques
progresses beyond simple naming, you must present the relevant intraverbal prompt (such
as What is it? or What
color is it?)Language
simultaneously
multiple exemplars
for Fostering
andwith
Cognitive
Ability of the
nonverbal stimulus (for example, red ball, red square, and red car) and appropriate nonexemplars (for example, blue ball, pink ball, orange ball) to ensure that the relevant stimulus
his book offers a series of revolutionary intervention programs for applied
feature will control responding. See the Tact Training: Names, Features, and Actions
work inofhuman
and cognition
with autism
program for a sample
specificlanguage
instructions,
though youtargeted
will needat tostudents
create other
specificand
programs other
for specific
tact targets.disabilities.
ProgramsIt forpresents
advanced
tacts (such
prepositions
or
developmental
a program
drawnasfrom
derived stimutacts of sensory
experiences)
can
be
found
at
the
end
of
this
chapter
(see
Tact
Training:
lus relations that you can use to help students of all ages acquire foundational and
Prepositions and Tact Training: Sensory Experiences). These programs often work well
advanced
verbal,trialssocial,
and cognitivewithskills.
when receptive
program
are intermixed
tact trials (see chapter 7 for additional
information). See the Tact Training Data Sheet for a model data sheet for all tact
programs. The first part of this book provides step-by-step instructions for helping students
Tact responses
to intraverbal
(such verbal
as What
is that?andordevelop
What other
color isbasic
learn relationally,
acquireprompts
rudimentary
operants,
that?) constitute a functional repertoire; however, do not expect tacts taught exclusively
skills. Inprompts
the second
sectionas ofindependent
this book, descriptions
youll find ways
enhance
as responseslanguage
to intraverbal
to occur
of theto environment (for
example,
Thereand
s your
red ball)repertoires
in conversations.
Conduct
students
receptive
expressive
by developing
theirtraining
ability to(seeread,
the Play-Based
Tact Training:
NaturalandLanguage
Paradigm
and out
datahow
sheet)to inteach
spell, construct
sentences,
use grammar.
Finally,program
youll find
play situations with natural change agents (for example, parents) to facilitate this use of
students&toLeBlanc,
apply the2007;
skillsKoegel
theyveet al.,
learned
higherCharlop,
order cognitive
and social
language (Gillett
1987;toLaski,
& Schreibman,
1988). functions, including perspective-taking, empathy, mathematical reasoning, intelProcedurally,
LanguageThisParadigm
(NLP) hasanalytic
severaltraining
steps (Charlop-Christy
ligence,Natural
and creativity.
applied behavior
approach will help
et al., 1999; Gillett & LeBlanc, 2007; LeBlanc et al., 2006). Adult and child face each
make many
language
and cognition
other withstudents
fun, common
objectssubstantial
and toys and
(suchlasting
as cup,gains
fish,inboat,
and car).
The adult not
traditional
interventions.
presents anpossible
array ofwithobjects
from which
the child may select (in other words, a preference assessment) and restricts item access after selection. The adult models an appropriate
play activity and spoken phrase (such as Fish swims) using the item. Echoic response
RUTHareANNE
PH.D.,
BCBA, isitem
an associate
professor
in the Rehabilitation
approximations
shapedREHFELDT,
by providing
contingent
access with
continued
models
Institute
of
Southern
Illinois
University,
Carbondale.
She
holds
doctoral
and
degrees
of the descriptive phrase while the child plays with the toy for twenty to thirtymasters
seconds.
in
psychology
from
the
University
of
Nevada
and
a
bachelors
degree
in
psychology
from
The adult retrieves the item and presents a different exemplar by modeling a different the
University of Puget Sound. She is also a board-certified behavior analyst.
phrase for the same object (for example, blue fish) or presenting a new stimulus array
for selection.
The adult inserts a five-second delayed echoic prompt while modeling the
YVONNE BARNES-HOLMES, PH.D., is a lecturer in psychology in the department of psyaction in order
the opportunity
responses
withoutIreland.
any direct vocal model. See
chologyto atallow
the National
University for
of Ireland
in Maynooth,
the Play-Based Tact Training: Natural Language Paradigm program for instructions
and createForeword
your ownwriter
user-friendly
sheetPH.D.,
that allows
you tooftrack
targets
and whether
STEVEN C.data
HAYES,
is a University
Nevada
Foundation
Professor of
responses occur
beforeat theor after
echoicof Nevada,
prompts.Reno. He is author of innumerable books and scientific
Psychology
University
articles,during
including
successful
and tocommitment
therapyandworkbook
Out of Your
Responses
initialthetrials
occuracceptance
in response
a vocal model
the EOGetcreated
and Into(inYourother
Life. words, echoic and mand response features). As the procedure
by restrictedMindaccess
progresses and the five-second delay begins to occur prior to the ISBN:
vocal978-1-57224-536-5
model in response
to restricted access and the nonverbal stimulus, the mand and tact features of the childs
response become more evident. The therapist is advised to begin incorporating neutral5 7 to9 9 5
only mildly preferred stimuli over time with the goal of minimizing the EO control over
responding to ensure that the ultimate response form is a pure tact (and not an impure
mand or partialnewharbingerpublications,
tact). The responseinc.requirement can be shifted over time in several ways
www.newharbinger.com
so that children learn to make longer responses (in other words,9 781572
two- or245365
three-word

Applications for Learners with Autism and Other Developmental Disabilities

US $79.95

the amount of nonfunctional language emitted by children with autism (Karmali, Greer,
Nuzzolo-Gomez, Ross, & Rivera-Valdes, 2005).
Several basic tact targets should be distinguished and ordered with respect to curriculum. Initial training should focus on naming of familiar three-dimensional objects (for
example, toys, animals, clothing, foods, and body parts) and people (for example, teacher
and siblings), followed by naming of two-dimensional representations of those objects
and people. Subsequent targets include features or aspects of familiar objects (such as size,
color, and shape), followed by specific location types (for example, kitchen or playground).
More advanced tact targets include actions (such as jumping, rolling, or flying) and functions and classes followed by relational tacts including prepositions (for example, in/out
of the box, above/below, and in front of/behind) and relational descriptors (big/little, less/
more, or slow/fast).

Derived
Relational
Responding

PRACTICAL

APPLICATIONS
OF THE MOST
CURRENT
RESEARCH

Applications for Learners with Autism

and Other Developmental Disabilities

A
Progressive
Guide to
Change

Three sections of skill-building

strategies to help individuals with
developmental disabilities:

2 Establish prerequisite skills for normal language

2 Acquire early relational operants
2 Speak with meaning & listen with understanding
2 Read with comprehension
2 Use appropriate syntax
2 Develop reasoning, problem solving & creativity
2 Be more empathetic, observant & self-regulating

EDITED BY RUTH ANNE REHFELDT, PH.D., BCBA

& YVONNE BARNES-HOLMES, PH.D.
FOREWORD BY STEVEN C. HAYES, PH.D.

Derived Relational Responding

Applications for Learners
with Autism and Other
Developmental Disabilities
A Progressive Guide to Change

EDITED BY
RUTH ANNE REHFELDT, PH.D., BCBA
& YVONNE BARNES-HOLMES, PH.D.

Context Press
New Harbinger Publications, Inc.

Shaping. Shaping is defined as differential reinforcement of successive approximations or

providing reinforcers for behaviors that closely resemble the behavior that the child needs
to learn (Baum, 2005). Thus, in teaching the response bubble, an instructor might first
reinforce the sound buh, bah, or even beh for several trials while listening closely for
variable responses, reinforcing sounds that even more closely resemble the word bubble
(such as bub and buhl), and placing prior responses on extinction (in other words, no
longer providing reinforcement). This process continues until the childs response closely
matches the target response, in this case, bubble.
Harris (1975) discussed the use of shaping to teach language to nonverbal children.
She said that the five steps in using shaping to teach language are the following: First,
ensure that the child is attending to the teacher, and that the child has the ability to
sustain this attention for a short period of time. Second, Harris suggests that the therapist
begin teaching by reinforcing all of the childs vocalizations. Third, the therapist should
reinforce only vocalizations that occur within six seconds of the model. Fourth, the therapist should reinforce all vocalizations that occur within six seconds of the model that
resemble the model. Finally, the therapist should provide a novel model and repeat the
steps described previously. The therapist should continue to quiz the child on previously
learned words to make sure his or her skills on the newly learned word do not decrease
because they are not being used.
Transfer of stimulus control. All of the procedures listed above can be enhanced with
the introduction of prompts to occasion the target response or a close approximation.
When prompts are used, stimulus control over the behavior is established with the bridge
stimulus (echoic prompt, picture prompt, and so on) and stimulus control is gradually
transferred to the relevant EO as those prompts are removed. Selection of the prompt
type should be based on evaluation of the strength of other verbal operants, with strong
operants used to support the emergence of weaker operants. Children with strong echoic
repertoires (in other words, they display ready responses to a variety of vocal models)
would benefit from echoic prompts. For example, a teacher might use incidental teaching
to capture a naturally occurring EO of interest in a nearby sink to instigate teaching trials
for water (turning it on, playing, or drinking) by providing the relevant vocal models at
the right times (in other words, saying on or water while standing with his or her
hand on the faucet handle, or saying drink while water is running and the teacher is
holding a glass). Children with strong tact repertoires (in other words, they display ready
responses to an object or picture) would benefit from tact prompts (Arntzen & Almas,
2002). For example, a therapist might prompt a child who can tact a glass of water (saying
water when a glass of water is present) in the presence of the EO by giving him or her
salty foods, presenting the glass of water, and responding to the childs tact water with
a glass of water to establish reduction of thirst as the maintaining reinforcer, and thirst as
the controlling antecedent stimulus.
After several successful prompted trials, stimulus control can be transferred by delaying the vocal model (echoic prompt) or altering the prompt (for example, the tact prompt
changes from a picture of a glass to a faint outline of a glass, or a partial echoic prompt).
A commonly used strategy for transferring stimulus control with children with autism is
institution of a time delay before the presentation of the prompt (Charlop, Schriebman,
& Thibodeau, 1985). After successfully establishing stimulus control with the prompt,
the therapist pauses before the presentation of the vocal model or prompt to allow an

opportunity for the child to respond to the EO in isolation, rather than the combined EO
and echoic or tact prompt condition used in previous trials, often termed a spontaneous
request. Time delay can be implemented as a fixed delay (for example, five or ten seconds
on all trials) or a gradually increasing delay across trials (for example, one second, two
seconds, four seconds, six seconds) until the child responds during the delay interval.
Fixed time delay is favored for mand training with visible, tangible stimuli and after
many mands have been mastered; graduated time delay is recommended for initial mand
training (Charlop-Christy, LeBlanc, & Carpenter, 1999).

The Tact: Importance and Types

In common vernacular, tacts are labels or descriptions emitted by the speaker. Skinner
defined a tact as a response that is evoked by a particular object or event or property of
an object or event (1957, p. 82). Thus, a tact is occasioned by the presence of a discriminative stimulus and is maintained by social reinforcers (such as praise or continued interaction with the listener). For example, in the presence of an apple, the responses apple
or fruit would be praised, while the response orange would likely be corrected (No,
thats an apple!).
The tact differs from the mand in that the controlling antecedent is a nonverbal
stimulus rather than an EO, and the reinforcer is a general social one rather than specific
access to a preferred event or stimulus change. For example, a child who sees an apple
and says apple in response to a parent query What do you call this? has emitted a
tact. The hungry child who says apple and immediately consumes the provided snack
has manded for the apple. Unlike mands, which are directly beneficial to the speaker,
tacts are primarily beneficial for the listener, as they provide descriptive information that
may enhance understanding or provide specificity. A child who tells a peer, Look in the
tree! (mand) may provide additional specificity by describing, That black kitty is going
to jump (tact). The listener now knows exactly what to attend to and can respond in a
social interchange, Thats too high. It might get hurt! Unfortunately, generalized social
reinforcers are notoriously unmotivating for young children with autism compared to
other types of reinforcers (in other words, tangibles); however, tacts are a basic building
block for conversation and are critical for success in general education environments.
The tact also differs from listener behavior, which is often referred to by various terms
including receptive discrimination, receptive labeling, or receptive vocabulary. Tacts refer to
the behavior of the speaker (in other words, naming or describing a critical aspect of the
environment), while listener behavior refers to ones ability to respond in various ways to
the important aspects of the environment specified by others. In the example provided
above, the speaker is directing a listener to observe a specific aspect of the environment,
the black kitty in the tree (tact), while the selection response of touching the picture of
the black kitty in an array of pictures of different animals would be a receptive discrimination (listener behavior). These two repertoires are related but functionally distinct and
will not necessarily develop without the direct targeting of each. Tacts are beneficial in
that they can facilitate social interactions, they facilitate development of generalized conditioned reinforcers by pairing weak social reinforcers (such as praise) with potent preferred items (Partington, Sundberg, Newhouse, & Spengler, 1994), and they may reduce

Procedures for Teaching Mands

The goal of mand training is to teach a child to request an item or activity in response
to the relevant EO without a prompt (for example, What do you want?). This can be
accomplished by ensuring that the relevant EO is present during learning trials (Sundberg
et al., 2002).

EO Manipulations
As stated previously, mands taught in the presence of an EO are more likely to occur
whenever this EO is present, even if the EO is present in novel environments or with a
variety of people (LeBlanc, Esch, Sidener, & Firth, 2006). Two primary strategies for
ensuring that training occurs in the presence of the EO are to capture naturally occurring
EOs for learning opportunities, and to create the EO as a specific part of the instructional
plan (Shafer, 1994). Most practitioners use both of these approaches at different points in
a childs day.

Captured EOs: Incidental teaching. Several studies illustrate the beneficial effects of
incidental teaching with children with autism (Fenske, Krantz, & McClannahan, 2001;
McGee, Morrier, & Daly, 1999), in which the child initiates the learning trial as he
or she navigates the natural environment and the therapist monitors the childs interests
and creates related learning opportunities. Incidental teaching involves five steps (Hart
& Risley, 1980). First, allow the child to interact naturally with his or her environment.
Second, watch the child closely to identify an opportunity for the child to request a
desired or needed item in the environment. Third, respond to the opportunity by requiring that the child request the item in his or her communication modality (for example,
vocal, sign, PECS, or VOCA). Fourth, use shaping or transfer of stimulus control procedures (see below) to facilitate an appropriate response from the child if the child does
not immediately emit the appropriate mand. Finally, immediately give the requested item
to the child when he or she gives an appropriate response and, if necessary, praise the
child for emitting the request (for example, say, Good job asking for the cookie). As
the mand response becomes stronger, eliminate the praise and provide only the requested
item. See the Captured Establishing Operations: Incidental Teaching with Time Delay
program for a detailed list of steps for using incidental teaching to teach mands which
can be used with the Mands Data Sheet. A program for teaching mands for removal
of aversive stimuli can be found at the end of the chapter (see Capturing Establishing
Operations: Terminating Aversive Stimuli).
Contriving EOs. Another way to ensure the presence of the relevant EO during mand
training is to intentionally create or contrive situations in which an item or event becomes
momentarily highly reinforcing immediately before you prompt responding. Contriving
an EO essentially consists of tripping a child up so that the child must request an
item or event to ensure that a preferred or needed activity or event can occur. Similar to
incidental teaching, the therapist or parent has to be continually vigilant in watching for
situations that would make teaching opportunities. However, contriving EOs differs from
incidental teaching in that the parent or therapist is arranging the environment to occasion the trial rather than relying on naturally occurring situations to occasion the childs

interest. One might contrive thirst as an EO by providing salty foods (thereby momentarily increasing the reinforcing properties of a drink) and blocking access to a drink until
the child asks for it. One might also cover a light switch with a hand as the child enters
a room in order to increase his or her motivation to request that the light be turned on.
Placing preferred toys on a high shelf can set the occasion for mands as well. Finally, one
might hold a childs swing to block the forward progress until the child requests (either
with or without prompts), Let go! or Swing!
Several studies illustrate the beneficial effects of identifying behavioral chains and
contriving an EO by interrupting the chain at some point (Sigafoos, Kerr, Roberts, &
Couzens, 1994). Interrupted chain procedures typically begin with teaching a child to
complete a chain or targeting a previously learned chain (for example, eating cereal and
milk with a spoon, or cutting out a circle from construction paper). At some point in
the chain the instructor prevents access to a component required for the chain (such as
a spoon or scissors), creating an EO relevant to a mand trial for the relevant item. Note
that these interrupted chain procedures are particularly useful for testing and targeting
pure mands for missing items because the child requests the item when no visual cue for
the item is present.
There are four steps to the interrupted chain instructional procedure to teach mands
(Duker, Kraaykamp, & Visser, 1994). First, a behavioral chain requiring several steps
(such as making a sandwich or preparing and eating cereal) that the child can do independently should be identified. Second, a step of the chain should be identified as the
targeted mand (for example, Knife, please, when all other items needed to make the
sandwich are available, or Spoon, please, when all other components for serving and
eating breakfast cereal are available). Third, the EO is contrived by creating a learning
trial where the child has all the items needed to complete a task except for one item. In
the sandwich example, the child would sit at a table with peanut butter, jelly, bread, and
a plate, and the teacher or parent would say, Lets make a sandwich! Allow the child to
initiate the chain and potentially request the needed item independently before modeling the response or using some other transfer of stimulus control technique to facilitate
responding (see below). Finally, immediately provide the requested item upon request.
See the Contrived Establishing Operations: Interrupted Chains program for the steps
involved in using an interrupted chain procedure to teach mands and the Interrupted
Chains: Sample Chains list for potential chains. Use with the Mand Training Data
Sheet. A program for teaching mands for a missing item in a game-type context also can
be found at the end of this chapter (see Contrived Establishing Operations: The Whats
in the Bag? Program).

Supplemental Teaching Procedures

In addition to the EO manipulation procedures described above, general instructional
strategies such as shaping and transfer of stimulus control procedures should be incorporated to increase the effectiveness of mand training. Errorless learning procedures are also
recommended such that early trials have immediate prompts sufficient to produce the
desired response, which are subsequently delayed (in other words, time delay) or altered
(in other words, partial vocal prompt) and gradually eliminated. See Green (2001) for
a review of errorless learning and other stimulus control technology procedures. These
supplemental procedures are equally appropriate for tact training.

Publishers Note
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with
the understanding that the publisher is not engaged in rendering psychological, financial, legal, or other professional services. If expert
assistance or counseling is needed, the services of a competent professional should be sought.

Distributed in Canada by Raincoast Books

Copyright 2009 by Ruth Anne Rehfeldt and Yvonne Barnes-Holmes
New Harbinger Publications, Inc.
5674 Shattuck Avenue
Oakland, CA 94609
www.newharbinger.com
All Rights Reserved

Acquired by Catharine Sutker; Cover design by Amy Shoup;

Edited by Karen Stein; Text design by Tracy Carlson
A copublication of New Harbinger Publications and Noetic Books

Contents
Series Editor Letter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
FOREWORD
An Applied Behavioral Psychology of Language and Cognition. . . ix
Introduction and Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 1

PDF ISBN: 978-1-57224-874-8

The Library of Congress cataloged the print edition as:

Derived relational responding : applications for learners with autism and other developmental disabilities / edited by Ruth Anne
Rehfeldt and Yvonne Barnes-Holmes ; foreword by Steven C. Hayes.
p. ; cm.
Includes bibliographical references and index.
ISBN-13: 978-1-57224-536-5 (hardcover : alk. paper)
ISBN-10: 1-57224-536-0 (hardcover : alk. paper)
1. Autistic children--Rehabilitation. 2. Developmentally disabled children--Rehabilitation. I. Rehfeldt, Ruth Anne. II. BarnesHolmes, Yvonne.
[DNLM: 1. Autistic Disorder--rehabilitation. 2. Association Learning. 3. Conditioning, Operant. 4. Developmental Disabilities-rehabilitation. 5. Reinforcement (Psychology) WM 203.5 D598 2009]
RJ506.A9D457 2009
362.1989285882--dc22
2008052211

PART 1

Establishing the Prerequisites for Normal Language

CHAPTER 1
Reinforcer Identification Strategies and Teaching Learner
Readiness Skills. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Thomas S. Higbee, Utah State University

CHAPTER 2
The Assessment of Basic Learning Abilities (ABLA) and Its
Relation to the Development of Stimulus Relations in Persons
with Autism and Other Intellectual Disabilities. . . . . . . . . . . . . . . . 25
W. Larry Williams and Marianne L. Jackson, University of Nevada, Reno

CHAPTER 3
Observing Responses: Foundations of Higher-Order
Verbal Operants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

CHAPTER 8
Acquiring the Earliest Relational Operants: Coordination,
Difference, Opposition, Comparison, and Hierarchy. . . . . . . . . . . 149

Dolleen-Day Keohane and Jo Ann Pereira Delgado, Columbia University

Teachers College and CABAS; and R. Douglas Greer, Columbia University
Graduate School of Arts and Science Teachers College

Carmen Luciano, Miguel Rodrguez, Israel Maas, and Francisco Ruiz,

University of Almeria, Spain; Nicholas M. Berens, Center for Advanced
Learning, Reno; Sonsoles Valdivia-Salas, University of Albany, New York

CHAPTER 4
Joint Attention and Social Referencing in Infancy as
Precursors of Derived Relational Responding . . . . . . . . . . . . . . . . . 63

CHAPTER 9
Applying Relational Operants to Reading and Spelling. . . . . . . . . 171

Martha Pelez, Florida International University

CHAPTER 5
Establishing Mand and Tact Repertoires. . . . . . . . . . . . . . . . . . . . . 79
Linda A. LeBlanc and Courtney M. Dillon, Western Michigan University;
and Rachael A. Sautter, Y.A.L.E. School

Deisy G. de Souza, Julio C. de Rose, and Camila Domeniconi, Universidade

Federal de So Carlos, Brazil

CHAPTER 10
Syntax, Grammatical Transformation, and Productivity:
A Synthesis of Stimulus Sequences, Equivalence Classes,
and Contextual Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Harry A. Mackay, Northeastern University and Praxis Inc.; and Lanny Fields,
Queens College and the Graduate School of the City University of New York

PART 2

Speaking with Meaning and Listening with Understanding

CHAPTER 6
Nonrelational and Relational Instructional Control. . . . . . . . . . . . 111
Jonathan Tarbox, Center for Autism & Related Disorders; Rachel S. F.
Tarbox, Chicago School of Professional Psychology at Los Angeles; and
Denis OHora, National University of Ireland, Galway

CHAPTER 11
Extending Functional Communication Through
Relational Framing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Rocio Rosales and Ruth Anne Rehfeldt, Southern Illinois University

PART 3

Self, Reasoning, Problem Solving, and Creativity

CHAPTER 7
Naming and Frames of Coordination. . . . . . . . . . . . . . . . . . . . . . 129
Caio F. Miguel, California State University, Sacramento; and
Anna I. Petursdottir, Texas Christian University

Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

CHAPTER 12
Training Analogical Reasoning as Relational Responding. . . . . . . 257
Ian Stewart, National University of Ireland, Galway; Dermot Barnes-Holmes,
National University of Ireland, Maynooth; and Tim Weil, University of
Nevada, Reno

the child) and may result in idiosyncratic or incomplete signs (the sign equivalent of poor
articulation or volume control).
Individual abilities and deficits that have an impact on failure to develop the mand
repertoire should also guide selection of a communication modality and mand training
procedures (Tincani, 2004; Wraikat, Sundberg, & Michael, 1991; Bourret, Vollmer, &
Rapp, 2004). Children may fail to exhibit mands due to dearth of items with reinforcing
properties, lack of a meaningful trained mand response, limited reinforcement opportunities (for example, poor articulation, and so few viable listeners), or because EOs were not
present during training and do not occasion responding. Bourret and colleagues (2004)
found that mand training is more effective when the training procedures target specific
problems. For example, one might target a lack of response by selecting a topography
based on the childs existing repertoires; vocal sounds and verbal imitation may indicate that spoken language should be pursued, whereas limited vocal imitation repertoire
with good fine motor and motor imitation skills may suggest sign language, PECS, or
VOCA.

The Mand: Importance and Types

Development of a functional mand repertoire can decrease problem behavior in children
with autism (Durand & Merges, 2001) by replacing inappropriate behaviors (such as tantrums and aggression) with less effortful and more appropriate means to obtain the same
functional reinforcers (such as attention, specific tangibles, or a break from an aversive
stimulus). Thus, teaching children to mand early in their verbal training can allow them
some control over their environment (terminating aversive stimuli, for example) and may
decrease the occurrence of problem behavior if the vocal response that is taught is less
effortful to emit than the problem behavior that is maintained by the same functional
reinforcers.
For the purposes of this chapter, mands will be divided into simple mands and mands
for information. Initial language training should focus primarily on establishing a robust
repertoire of simple mands across different environments and with different people.
Simple mands are those basic requests for preferred items (such as toys, food, or
drink), people, changes in the environment (for example, Lets go outside), and termination of unpleasant situations (such as Stop, please or taking a break). These mands
should be targeted as soon as a clear preference or motive is identified, and targeted continuously until a strong and spontaneous repertoire exists across environments and with
multiple people. A spontaneous repertoire means that mands occur under EO control (for
example, hunger, thirst, or boredom) rather than as the result of a prompt from another
person (for example, What do you want?), such that the child is able to use language
to meet his or her needs when they occur rather than waiting for someone to notice his
or her distress.
As training continues with preferred items, mands for missing preferred items should
be periodically tested to ensure that the relevant EO (such as hunger), rather than a
nonverbal discriminative stimulus (such as the sight of food), is controlling the childs
requests. For example, requests for a meal that occur only when food is visible or can be

smelled and not when the child is hungry would not be particularly functional. Mands for
missing items should be gradually introduced by conducting trials with known mands in
the absence of the discriminative stimulus (in other words, the actual preferred item). For
example, a child who likes cookies and is hungry might be given a small visible portion of
a cookie upon request followed by an opportunity to mand for additional cookies that are
not visible. Certain manipulations can be employed in order to evoke a mand for an item
in the absence of the actual item. For example, placing the child in a contextually relevant
environment (for example, bringing the child to the kitchen and orienting him or her
toward the cabinet where the cookies are located), but keeping the actual item that is to
be requested out of sight, may evoke this mand. The PECS also includes a phase for training mands in the absence of the actual item. This occurs in phase two of training, where
the items and the communication book are moved farther away from the learning and
typically out of the direct line of vision (Bondy & Frost, 2002). Mands that occur when
an item is missing are considered pure mands because of the certainty one can have that
the EO, rather than a visible discriminative stimulus, is the relevant antecedent variable.
Mands for removal of aversive stimuli should be taught in naturally occurring contexts
as soon as the aversive stimuli can be readily identified and removed. Often practitioners
fail to recognize the importance of teaching a specific mand for a break, for help, or to
stop until problem behavior maintained by escape from aversive situations has been well
established and must be subsequently replaced. Often, one to two generally applicable
mands for removal of aversive stimuli will suffice across many situations, while a large of
number of specific mands for items are required to produce functional repertoires. Mands
for removal of aversive stimuli should be targeted exclusively through captured EO programs, as opposed to contrived EO programs, to avoid creation of unnecessary distress
and increased likelihood of the therapist becoming a conditioned aversive stimulus.
A mand for information is a subtype of mand (Sundberg, Loeb, Hale, & Eigenheer,
2002) that specifies certain information as the desired reinforcer (for example, Who has
my Elmo? What time is it? Where are my keys? or What is your name?). There
are often at least two pertinent reinforcers (Sundberg et al., 2002): the information itself
and the outcome that the information allows the person to achieve. The child benefits
from the information because it allows quicker and more efficient problem solving than
a trial-and-error strategy (for example, approaching each person in the room to look for
the doll) would provide.
A mand for information is more complex than a simple mand because it requires
a child to identify the necessary information and a person who might have it, and to
formulate a reasonable and understandable question. Some mands for information occur
in a purely social context and have only one reinforcer (in other words, the information itself). For example, Where did you go on vacation? produces information that
facilitates ongoing conversation and social connection but does not provide any other
tangible stimulus change in the environment. Mands for information should be targeted
when a child has an extensive tact and developing intraverbal repertoire. Initial training
should focus on mands for concrete and powerful secondary reinforcers in addition to
information (for example, Where is my favorite toy?), while mands for purely social
information should be targeted later because children with autism may not strongly value
social information or small talk.

an establishing operation (EO; Michael, 1988). An EO has been defined as a change in

the environment that momentarily increases or decreases the power of a reinforcer and
evokes behaviors that have previously produced access to that reinforcer (Michael, 1982).
For example, hunger increases the reinforcing value of food and evokes behaviors that
have produced access to food in the past (for example, asking Mom for lunch or cooking).
A tact is a label or description that occurs in the presence of a nonverbal discriminative
stimulus or condition that results in social reinforcers. For example, in the presence of
an apple, a child who says the word apple would be praised; if the child said the word
orange, he or she would not. Much of early language development consists of these two
verbal operants as children learn to ask for things that they want and describe things that
they see. Although they are not the focus of this chapter, other elementary verbal operants
such as echoics (or verbal imitation) and textual behavior (the term for reading), are sometimes used as prompts in training mands and tacts. These verbal operants are conceptualized as functionally independent such that a child who mands cookie when he wants
one may not be able to tact, or name a cookie when he sees a picture of one, a pattern that
has been repeatedly demonstrated to occur in children with language disorders (NuzzoloGomez & Greer, 2004; Sigafoos, Reichle, Doss, Hall, & Pettitt, 1990).

Applying Skinners Analysis of Language

to Instruction
Sundberg and Michael (2001) outline several benefits of using Skinners analysis of language to teach children with autism. First, Skinners identification of the functional independence of the verbal operants indicates that each one has to be independently and
directly targeted or it may not develop. In other words, children with autism can only be
expected to develop functional mands or requests if they are taught in the presence of an
important motivational variable (for example, learning to ask for water when thirsty) and
not just as a product of training to tact (label) a glass of water (Twyman, 1996). Sundberg
and Michael (2001) also identify the mand as the type of operant to teach first to children
with autism because the mand is uniquely valuable to the child and doesnt require social
reinforcers. That is, a child learns to mand for things he or she desires when the desire
exists, allowing the child some control over his or her environment (for example, I ask
for an item and receive it). In addition, the child receives a powerful reinforcer following a verbal response (in other words, the child receives what he or she asks for), which
can increase the reinforcing properties of speaking, resulting in a greater likelihood of
language development in the future.
When one is preparing to teach mands and tacts, it is critical to assess the childs
existing repertoires for two purposes. First, one must determine the current profile of
functional language in order to identify which operants should be targeted and to determine which existing operants can be used to prompt weaker or emerging operants.
Second, one must determine the extent of existing prerequisite repertoires, which could
affect your selection of topography of the target response. Each of these areas is covered
briefly below.

Functional Language Abilities

The Assessment of Basic Language and Learning Skills (ABLLS) is one tool based
on Skinners analysis that allows therapists to identify a childs specific language deficits
in terms of verbal operants (Partington & Sundberg, 1998). Another tool is the Verbal
Behavior Milestones Assessment and Placement Program (VB-MAPP; Sundberg & Hale,
2007). Each tool evaluates Skinners verbal operants to determine whether a child has a
specific verbal operant in his or her repertoire and allows you to track development via
repeated administration. Each verbal operant is assessed under the relevant antecedent
conditions and with the corresponding consequences in place to determine if the child
readily emits that verbal operant. The information obtained from the ABLLS can inform
parents and teachers about how to prioritize language targets. However, generally children should learn mand (requesting) and echoic (verbal imitation) skills before moving
on to tacts (labels), questions (mands for information), and conversation (intraverbals;
Sundberg & Partington, 1998). The obtained profile can also allow the teacher or parent
to identify strong repertoires (such as echoics) that can be used in teaching other operants
(such as mands). (See the section Transfer of stimulus control, below.)

Response Topography
Selecting a modality for responding is critically important to success and maintenance
in language instruction with children with autism. Language, or verbal behavior, does not
have to be vocal (in other words, spoken) to be meaningful or functional. Verbal behavior
can also include manual signing (Tincani, 2004; Bartman & Freeman, 2003), selection
responses such as picture exchange (Charlop-Christy, Carpenter, Le, LeBlanc, & Kellet,
2002; Chambers & Rehfeldt, 2003) or use of voice output devices (Mirenda, 2003). All
potential communication modalities should be considered for all children, with modality
selection based on three important factors: available audiences, practicality or portability
of the system, and child repertoires.
Vocal responses are ultimately portable and have broad potential listener communities,
assuming adequate articulation, but they have the drawback of requiring complex vocal
musculature manipulations that cannot be directly prompted. Picture exchange communication systems (PECS) and voice output communication aids (VOCA) can be used
with almost any listener community but must be portable, well-organized, and constantly
available. Researchers have proposed several drawbacks associated with selection-based
systems such as PECS. Sundberg and Michael (2001) distinguish between topographybased systems, where there is a unique topographical response (such as a spoken word
or sign) for each communicated idea, and selection-based systems, where the series of
responses (such as scanning, selection, and picture delivery) is identical for each communicative event. Sundberg and Partington (1998) argue that selection-based systems
often require more complex skills than initial appearance would suggest, with increased
probability of difficulties in acquisition as language concepts become more abstract and
difficult to depict visually. Sign language has the benefits of being topography based and
amenable to modeling and manual guidance, but it has the drawback of a restricted
verbal community (such as a deaf signing community and those trained to listen to

CHAPTER 13
Understanding and Training Perspective Taking as
Relational Responding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Louise McHugh, University of Wales, Swansea; and Yvonne Barnes-Holmes
and Dermot Barnes-Holmes, National University of Ireland, Maynooth

CHAPTER 14
Establishing Empathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Sonsoles Valdivia-Salas, University of Albany, New York; Carmen Luciano, University
of Almeria, Spain; Olga Gutirrez-Martinez, University of Central Barcelona, Spain;
and Carmelo Visdmine, Justice Administration, Madrid, Spain

CHAPTER 15
Mathematical Reasoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Chris Ninness, James Holland, Glen McCuller, Robin Rumph, Sharon Ninness,
and Jennifer McGinty, Stephen F. Austin State University; and Mark Dixon,
Southern Illinois University

CHAPTER 16
Developing Self-Directed Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . 335
Carmen Luciano, University of Almeria, Spain; Sonsoles Valdivia-Salas,
University of Albany, New York; Francisco Cabello-Luque, Universigy of
Murcia, Spain; and Monica Hernandez, Universigy of Jan, Spain

CHAPTER 17
Teaching Flexible, Intelligent, and Creative Behavior. . . . . . . . . . . 353
Catriona OToole, Carol Murphy, and Dermot Barnes-Holmes, National University
of Ireland, Maynooth; Jennifer OConnor, ABACAS, Kilbarrack, Ireland

Dear reader,
Welcome to New Harbinger Publications. New Harbinger is dedicated to publishing
books based on acceptance and commitment therapy (ACT) and its application to specific areas. New Harbinger has a long-standing reputation as a publisher of quality, wellresearched books for general and professional audiences.
Most existing books in the ACT series focus on teaching either practitioners or
members of the general public how to develop acceptance and mindfulness skills, and
how to apply those skills to various aspects of their experience. Although theoretical constructs and concepts are discussed to some extent in these books, they never take center
stage. This book is different. It explicitly deals with relational frame theory, a new behavioral psychology of language and cognition that has emerged on many fronts over the
past thirty years. Human behavior is qualitatively different from and more complex than
animal behavior, and many would agree that it is the complexity of human language and
cognition that sets us apart from nonhuman animals. So making advances in understanding how language and cognition work to our advantage and to our detriment is truly
important.
One of the goals of acceptance and commitment therapy is to help people move
beyond the tyranny of the mind that brings so much unnecessary suffering into the lives
of virtually all of usnot just people with clinical diagnoses. Because RFT provides an
understanding of how the mind works, it is helpful for an ACT therapist (and in fact
any other therapist) to know and understand RFT. However, psychotherapy is only one
application of RFT, and this book makes it clear that RFT has much broader practical
implications and applications. A glance at the table of contents will immediately show
the incredibly broad scope of RFT. Indeed, what sets this book apart from other books
on RFT is the formidable range of areas and applied topics from education and clinical
psychology that are covered in it. Several of the chapters examine specific skills that are
essential to all human functioning, such as reasoning (including mathematical reasoning),
perspective taking, and establishing empathy. All chapters present a balance of theory,
empirical data, and specific applications that bring theoretical concepts to lifeoften in
the form of real-life or case examples. The authors also offer useful suggestions on how to
apply RFT knowledge in a variety of contexts.
As part of New Harbingers commitment to publishing books based on sound, scientific, clinical research, we oversee all prospective books for the Acceptance and Commitment
Therapy Series. Serving as ACT series editors, we comment on proposals and offer guidance as needed, and use a gentle hand in making suggestions regarding the content,
depth, and scope of each book.
Books in the Acceptance and Commitment Therapy Series:
 Have an adequate database, appropriate to the strength of the claims
being made.

Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

 Are theoretically coherent. They will fit with the ACT model and underlying behavioral principles as they have evolved at the time of writing.
 Orient the reader toward unresolved empirical issues.
 Do not overlap needlessly with existing volumes.

 Avoid jargon and unnecessary entanglement with proprietary methods,

leaving ACT work open and available.

FOREWORD

 Keep the focus always on what is good for the reader.

 Support the further development of the field.
 Provide information in a way that is of practical use to readers.
These guidelines reflect the values of the broader ACT community. Youll see all of
them packed into this book. They are meant to ensure that professionals get information
that can truly be helpful, and that can further our ability to alleviate human suffering by
inviting creative practitioners into the process of developing, applying, and refining this
approach. Consider this book such an invitation.
Sincerely,
Georg H. Eifert, Ph.D., John Forsyth, Ph.D., and Steven C. Hayes, Ph.D.

An Applied Behavioral Psychology of

Language and Cognition
Behavior analysis is a unique tradition in psychology. It was based on the bold idea that
a set of analytic principles could be developed from work with nonhuman animals that
would eventually enable the understanding and analysis of human complexity. Studying
the relatively simple acts in relatively simple contexts of organisms with relatively wellknown histories was a strategy, not the final goal of the research. B. F. Skinner said his
approach was to create a promising conception of human behavior by a research focus
which began with simple organisms in simple situations and moved on, but only as its
growing power permitted, to the complexities of the world at large (Skinner, 1938 p. xiv).
Behavioral psychology aspired to give a comprehensive account of psychological phenomena. It was just taking a bottom-up approach to getting there.
Behavior analysts did not know if this strategy would succeed. It was not a certainty
it was a hope. There is nothing in evolutionary processes that ensures that the behavior of
nonhumans must provide a good guide to the analysis of complex human behavior.
Remarkably, it often does. Indeed, the behavior analytic strategy worked almost too
well. Functional principles of behavior proved to be powerful tools with vast applied
relevance, but they did not seem to provide a fully adequate set of tools for the empirical analysis of language and cognition. That combination was a bad one for behavioral
psychology. Progress was good enough to keep adherents happy but poor enough to lead
others to abandon hope of ever succeeding.
By the mid 1970s the mainstream had moved elsewhere, convinced that the original
dream of a comprehensive psychology was either unattainable or unlikely to be successful within the behavioral tradition. Psychology focused on cognitive science. Clinicians
moved on to clinical theories of cognition. Microtheories were the norm, and grand
visions fell out of favor.
And there it stood for a few decades. But in the last fifteen years something remarkable has occurred. A true behavioral psychology of language and cognition has begun to
form. Initially it resided almost entirely in the basic behavioral laboratories, but ever so
gradually it has emerged in the applied arena as well. As befitted the inductive behavior
analytic tradition, each step built on the work of the past. Little needed to be thrown
away; instead, more and more was added until, finally, it became reasonable to suppose
that behavioral psychology could successfully and empirically move on, as its growing
power permitted, to the complexities of human language and cognition.
One began to see the implications of this shift first in clinical behavior analysis as
approaches such as acceptance and commitment therapy, dialectical behavior therapy, and
functional analytic psychotherapy began to revitalize the behavioral wing of cognitive

Pelez, M., Lubin, J., McIlvane, W., & Dube, W. (2001). Training discrimination, reflexivity, mixed identity matching, generalized identity matching, and arbitrary matching in
infants: Towards the emergence of stimulus equivalence. Presented at the first international conference of the Association for Behavior Analysis, Venice, Italy.
Pelez-Nogueras, M., Field, T., Hossain, Z., & Pickens, J. (1996). Depressed mothers
touch increases infant positive affect and attention in still-face interactions. Child
Development, 67, 17801792.
Pelez-Nogueras, M., & Gewirtz, J. (1997). The context of stimulus control in behavior
analysis. In D. M. Baer & E. M. Pinkston (Eds.), Environment and behavior. Boulder,
CO: Westview Press.

CHAPTER 5
Establishing Mand and Tact Repertoires
Linda A. LeBlanc and Courtney M. Dillon, Western Michigan
University; and Rachael A. Sautter, Y.A.L.E. School

Children with autism typically exhibit substantial speech delays during early childhood,
with up to 50 percent failing to develop any speech (Charlop & Haymes, 1994). Those
who do develop speech often engage in echolalia, repeating words or phrases heard previously, or do not speak for social purposes such as engaging in conversation (Smith, 1999).
Because one of the best predictors of outcome for children with autism is the development of spontaneous language before six years of age (Szatmari, Bryson, Boyle, Streiner,
& Duku, 2003), it is essential to develop effective programs to teach language to children
with autism. This chapter describes behavioral techniques that can be used to teach basic
language to young children with autism in accordance with B. F. Skinners analysis of
language.

Skinners Analysis of Verbal Behavior

As discussed in chapter 3 of this volume, Skinner (1957) defined language or verbal
behavior as behavior reinforced through the responses provided by other people, and he
defined several verbal operants according to the relevant antecedents and consequences.
He classified language according to specific functional operants (for example, antecedentbehavior-consequence groups), rather than the traditional units of language (such as words
and phrases). This chapter will focus on two of Skinners elementary verbal operants, the
mand and the tact. In common terms, a mand is a request (such as May I have that?)
or an imperative (such as Give it to me). The mand tells the listener what response or
item will serve as a reinforcer, and it occurs because of a highly specific motivation called

higher cognition. For example, it seems likely that the emotional and social aspects of
social referencing form the basis of the later development of perspective taking. That
is, reciprocal conversation, cooperative play, and displays of sympathy and empathy for
others are all social abilities that require the basics of joint attention and social referencing, because without them you would not use the ongoing cues of others to determine
how they were feeling and to act accordingly. Hence, it is not surprising that individuals
with autism who present with deficits in social referencing, for example, subsequently
develop considerable delays in their social and emotional skills.

Concluding Comments
Joint attention and social referencing are an intricate part of the tapestry of social interactions that comprise normal development. Not only are they critical to the development
of social and related emotional repertoires, but they also appear to be essential precursors
to conditional discriminations and identity matching, which are also important precursors to language development and its core process of derived relational responding. The
current chapter described teaching strategies for establishing conditional discriminations,
joint attention, and social referencing in young and developmentally disabled learners.
Despite the importance of these skills, such training is far from easy. But there is simply
no way around thisif language and social and emotional development are desired and
potentially within the capabilities of the learner, then the difficulties must be endured
and the teacher must generate increasingly clever and creative ways to make the training work. Although empirical evidence in support of the various teaching strategies outlined is still scant, they offer good first steps toward the establishment of these essential
building blocks of human development.

Devany, J. M., Hayes, S. C., & Nelson, R. O. (1986). Equivalence class formation in
language-able and language-disabled children. Journal of the Experimental Analysis of
Behavior, 46, 243257.
Dube, W. V., MacDonald, P., Mansfield, R. C., Holcomb, W. L., & Ahearn, W. H. (2004).
Toward a behavioral analysis of joint attention. Behavior Analyst, 28, 197207.
Feinman, S. (1982). Social referencing in infancy. Merrill Palmer Quarterly, 28,
445470.
Gewirtz, J. L., & Pelez-Nogueras, M. (1992). Infant social referencing as a learned
process. In S. Feinman (Ed.), Social referencing and the social construction of reality in
infancy. New York: Plenum.
Gewirtz, J. L., & Pelez-Nogueras, M. (2000). Infant emotions under the positivereinforcer control of caregiver attention and touch. In J. C. Leslie & D. Blackman
(Eds.), Issues in experimental and applied analyses of human behavior. Reno, NV:
Context Press.
Higbee, T. S., & Pelez-Nogueras, M. (1998). Reinforcer identification in infants.
Behavioral Development Bulletin, 7, 1014.
Holth, P. (2005). An operant analysis of joint attention skills. Journal of Early and Intensive
Behavioral Interventions, 2, 160175.
Jones, E. A., Carr, E. G., & Feeley, K. M. (2006). Multiple effects of joint attention intervention for children with autism. Behavior Modification, 30, 782834.
Klinnert, M., Campos, J. J., Sorce, J. F., Emde, R. N., & Svejda, M. (1983). Social referencing: Emotional expressions as behavior regulators in emotion. Theory, Research,
and Experience, 2, 5786.

References

MacDonald, R., Anderson, J., Dube, W. V., Geckeler, A., Green, G., Holcomb, W.,
et al. (2006). Behavioral assessment of joint attention: A methodological report.
Developmental Disabilities, 27, 138150.

Augustson, K. G., & Dougher, M. J. (1992). Teaching conditional discrimination to

young children. Experimental Analysis of Human Behavior Bulletin, 9, 2124.

McClannahan, L. E., & Krantz, P. J. (2006). Teaching conversation to children with autism:
Scripts and script fading. Bethesda, MD: Woodbine House.

Bakeman, R., & Adamson, L. (1984). Coordinating attention to people and objects in
mother-infant and peer-infant interaction. Child Development, 55, 12781289.

Mundy, P., Sigman, M., & Kasari, C. (1994). Joint attention, developmental level and
symptom presentation in autism. Development and Psychopathology, 6, 389401.

Carpenter, M., Pennington, B. F., & Rogers, S. J. (2002). Interrelations among socialcognitive skills in young children with autism. Journal of Autism and Developmental
Disorders, 32, 91106.

Mundy, P., Sigman, M. D., Ungerer, J., & Sherman, T. (1986). Defining the social deficits of autism. The contribution of non-verbal communication measures. Journal of
Child Psychology and Psychiatry and Allied Disciplines, 27, 657669.

Charman, T., Swettenham, J., Baron-Cohen, S., Cox, A., Baird, G., & Drew, A. (1997).
Infants with autism: An investigation of empathy, pretend play, joint attention, and
imitation. Developmental Psychology, 33, 781789.

Pelez, M., Gewirtz, J. L., Sanchez, A., & Mahabir, N. M. (2000). Exploring stimulus
equivalence formation in infants. Behavior Development Bulletin, 9, 2025.

Dawson, G., Toth, K., Abbott, R., Osterling, J., Munson, J., Estes, A., et al. (2004). Early
social attention impairments in autism: Social orienting, joint attention, and attention to distress. Developmental Psychology, 40, 271283.

Pelez, M., Gewirtz, J. L., & Wong, S. E. (2007). A critique of stage theories of human
development: A pragmatic approach in social work. In B. A. Thyer (Ed.), Comprehensive
handbook of social work and social welfare: Vol. 2. Human behavior in the social environment. New York: Wiley and Sons.

behavioral therapy. But these clinical extensions were not yet point-to-point empirical
extensions of the new basic behavioral work in language and cognition. That would take
more time.
This book is the next giant step in that process. It is the first to demonstrate a comprehensive set of applied behavior analytic training approaches for language and cognition that directly addresses most of the key areas within that domain. The chapters avoid
needless quarrels between competing factions within basic behavior analysis; theories in
this volume are treated more as useful tools than as distinctions between warring camps.
Even if individual chapters largely adopt a particular perspective, considered as an entire
set they give testimony to the emergence of an increasingly unified behavior analytic
approach that is now ready to walk, step-by-step, from the simplest learning tasks all the
way through empathy, self, and creativity.
That is a notable achievement, and one that may be a first in applied psychology. I
know of no other book that extends a single approach within basic experimental psychology into intervention programs across the full range of issues that need to be addressed in
applied work in human language and cognition.
This book, intended for parents and a variety of professionals working with individuals with autism and other developmental disabilities, who understand the principles of
applied behavior analysis, begins with the earliest steps needed to establish the prerequisites for normal language, helping applied workers think through how to determine
reinforcers; how to establish observation, attention, and simple discriminations; and how
to establish a simple repertoire of Skinnerian verbal operants. That section is closest to
traditional applied behavior analysis and it deals with issues that many cognitive scientists
do not address because they deal with higher-functioning participants in their studies of
reading or problem solving. The book then steps firmly and boldly into derived relational
responding and the issues of meaning and understanding. Chapters address instructional
control, naming, acquisition of relational framing, syntax, reading, and functional communication. These chapters represent important steps forward within behavior analysis,
bringing together research that is reasonably well-known but also showing in the totality how much progress has been made. Finally, the book dances into some of the most
complex issues of all as it considers self, reasoning, problem solving, and creativity, with
chapters on analogy, perspective taking, empathy, self-rules, mathematics, and creativity.
Especially in this last section, the volume reveals how bold and cutting-edge it is. Yet even
in this final section, and throughout the entire book, all of the chapters have empirical
support.
What is almost startling is that this book is not mere interpretation and logical extension, or a broad vision to be tested in some distant future. It is also not a volume that tries
to declare by fiat that a limited empirical analysis obviates an analysis of more complex
issues. We have seen such volumes before. What we have not seen before is a comprehensive empirical book that covers the full range of applied topics that educators and
clinicians can begin to use now. In its scope, practicality, and empirical base, this volume
declares that a comprehensive applied behavioral psychology of language and cognition is
here, is real, and is moving ahead.
Will this approach fully succeed as measured against the bold goals of behavior analysis? That we do not know. But a volume like this is needed to help applied workers
take the steps to find out; within the inductive, technique-building tradition of behavior

analysis, each step forward is likely to create progress that is sustained, since even when
well-crafted steps fail, they provide important information.
Applied behavior analysis is a sophisticated and vigorous area with thousands of
sophisticated and creative applied professionals. I cant wait to see what all of the wonderful behavioral educators and practitioners do with this bold new approach.
Steven C. Hayes
University of Nevada

References
Skinner, B. F. (1938). Behavior of Organisms. New York: Appelton-Century-Crofts.

What this research demonstrates is that the extent to which an infant orients to the
mothers face for cues in contexts of uncertainty depends on past success in obtaining
such information, its validity, and its utility. For training purposes, therefore, either with
very young learners or with those who are developmentally disabled, the cues of others
must be consistently contingent on the learners object-referencing behavior and must
reliably predict environmental consequences for the learners approach or avoidance. This
interpretation is summarized in Figure 4.5.

Introduction and Acknowledgments

In the 1970s, Murray Sidman and colleagues made an important, albeit unplanned, discovery that would serve as the impetus for a prosperous research program in behavior
analysis for years to come. Sidman (1971, 1977, Sidman & Cresson, 1973) found that,
after researchers explicitly taught individuals with developmental disorders and limited
language skills to match dictated names to the corresponding pictures and the pictures to
their corresponding printed words, the individuals proved capable of naming the pictures,
orally reading the text, and matching words to pictures and pictures to wordsa skill
seemingly indicative of reading comprehensionall in the absence of direct instruction.
Sidman (1971, 1977) termed these novel, emergent relations equivalence relations, as the
untrained skills seemed to represent symbolic or referential behavior, phenomena for which
a behavior analytic explanation had previously been lacking. These findings, and those
that would follow in years to come, were exciting for the field of behavior analysis, for
they identified an economic and efficient means of establishing new academic and pre
academic repertoires in individuals with significant learning challenges, as each desired
skill did not have to be directly taught. But what was perhaps even more important is that
these developments, even where they were academic, have led to significant transformations in our understanding of the basic processes of human language and cognition
making it possible that in years to come there will be no areas of learning that cannot be
taught.
In the years since Sidmans pioneering discoveries, a voluminous body of laboratory
research has been conducted that further elucidated the conditions necessary and sufficient for the establishment of stimulus equivalence, multiple stimulus relations, and
indeed whole networks of derived or untrained stimulus relations. This research led to
the formulation of several major theories regarding the nature of relational learning (such
as relational frame theory and the naming hypothesis), which have inspired even more
sophisticated types of experimental research and innovative methodologies. Of course,
it is not of paramount importance that we determine which is the best theory at one
level or another. But the best test is one that shows which theory delivers most into the
lives of those who need it, either scientifically and indirectly in the form of the advancement of knowledge or directly in the form of the development of new educational technologies. And on both counts we have much to inspire us. For example, we know that
establishing a history of reinforced relational responding in individuals with a variety
of learning challenges is an effective and efficient means of programming for the emergence of such educationally relevant skills as reading and spelling (Hanna, de Souza, de
Rose, & Fonseca, 2004), recognizing names and faces of caregivers (Cowley, Green, &
Braunling-McMorrow, 1992), requesting preferred items (Rosales & Rehfeldt, 2007), and
understanding basic numerical concepts (Lynch & Cuvo, 1995), to name a few. Thus,

Figure 4.3. Mother signals joyful cue to infant that predicts pleasant musical sound,
contingent on infant reaching for object.

Figure 4.5. A learning approach to a social referencing paradigm.

From this perspective, social referencing is an example of social knowledge with an
emotional component. Put simply, the experienced learner gains the knowledge that if
another is smiling when a stranger approaches, reinforcement is likely; but if the other
person is cringing, for example, reinforcement is not likely. Thus, the facial expression
of the other becomes a setting event that establishes the function of the stranger as
being discriminative for positive or negative reinforcement or aversive consequences for
approaching. In line with this analysis, it should then be possible to establish learners
responses to the basic emotions displayed by others and how they should act on this basis.
Once these have been established, it is likely that a whole array of more subtle emotional
reactions and appropriate response patterns will be trainable within the context of simple
conditioning paradigms.

Why Are Joint Attention and Social Referencing

Important for Derived Relational Responding?
Figure 4.4. Mother signals fearful cue to infant that predicts loud sound and movement
of object, contingent on infant reaching for object.

Joint attention and social referencing would seem to have an important role in the
establishment of derived relational responding, thus forming the core of language and

a visual path from your eyes to the treat as you place it under a cup. This can also be
trained explicitly. This is to be repeated until the learner chooses the right cup.

Joint object attention. With joint object attention, the learner orients quickly or directly
toward an object once another persons attention to the object has been discriminated. At
best, the learner should also initiate your attention once a novel object has been identified.
Consider the following scenario described by Jones, Carr, and Feeley (2006). Position a
toy of choice less than five feet away from the learner, activate the toy, and turn and look
at the learner while pointing to the toy and commenting upon it (for example, Look at
what the car is doing). It should be possible to get the learner to orient toward the toy
within as little as two seconds of your comment. Then, to improve initiation on behalf of
the learner, have him attend to the toy for several seconds, and encourage him to point
to the object while he looks at you. To reinforce this pointing response, you can simply
model it or physically form the learners hand to point at the object.
Mutual object orienting with gestures. Holth (2005) described the following steps for
establishing mutual object orienting with gestures. Attach five or six envelopes to a wall
in a horizontal line. In view of the learner, who is seated approximately ten to thirteen
feet away from you, place an edible reinforcer in one of the envelopes. In order to access
the snack, the learner must guide you through the envelopes. That is, you will begin by
pointing to the envelope farthest away from the one containing the snack, and prompt
the learner to guide you with simple directives such as left and right, and stop when
you reach the correct envelope. You can also arrange the envelopes in a vertical line and
include prompts such as up and down. Ultimately, you should be able to arrange the
envelopes in a semirandom sequence (some side-by-side and others above and below) and
all direct prompts to the learner should be faded. It is interesting to note some empirical
evidence suggests that this type of intervention not only improves mutual orienting and
gesturing, but is also associated with language gains (Jones et al., 2006).

Empirical Evidence
There is some empirical evidence to support the use of training regimes for the establishment of joint attention in children with autism. In one study, MacDonald and colleagues (2006) investigated joint attention initiations in twenty-one typically developing
children (ages two to four) and twenty-six children with autism. As expected, the children
with autism demonstrated relatively minor deficits in joint attention responding and more
severe deficits in joint attention initiation. While the majority (78 percent) demonstrated
gaze shifts, 44 percent demonstrated use of gestures, and only 22 percent were capable of
related vocalizations. However, after one year of participation in a comprehensive treatment program, all of the children with autism demonstrated gaze shifts, all had gestures,
89 percent could vocalize, and levels of joint attention were now commensurate with the
normally developing counterparts.
A study by McClannahan and Krantz (2006) also demonstrated the remediation of
deficits in joint attention in three children with autism (ages two to five). In this research,
photographic activity schedules were used to cue learners to play with toys in three
locations a puppet theater, toy shelves, and a toy box. When learners initiated use of toys,

they were manually guided to point to the toy while orienting to the teacher. Across trials,
manual prompts were faded (from graduated guidance to spatial fading and shadowing),
and the teachers proximity was decreased gradually. The results indicated that all three
children learned to point and orient for attention and could do so with novel stimuli.

The Concept of Social Referencing

Although numerous authors integrate the concepts of joint attention and social referencing, the current chapter argues that they are distinct and that joint attention essentially
precedes social referencing. Specifically, what social referencing adds to joint attention is
that it also involves the learner reacting to the novel stimulus in a manner that is in accordance with the others expression (Pelez-Nogueras & Gewirtz, 1997). Consider again the
previous example of the three-year-olds surprise when the kitten runs into the room. As
part of her joint attention skills, the child looks up at her mother while pointing to the
kitten, but then she engages in social referencing when she sees her mother make a fearful
face and as a result avoids approaching the kitten.
As well as incorporating the component of concordant responding (in other words,
using the reactions of others as discriminative stimuli for ones own responding), social
referencing also appears to comprise an emotional component. In other words, it extends
beyond the simple sharing of information and also facilitates the learners emotional reaction to stimuli. This emotional aspect of social referencing appears to make up a four-stage
process that involves recognizing emotional expressions, understanding emotional expressions, responding to emotional expressions as cues, and altering behavior in accordance
with changes in emotional expression.
Cognitive-developmental psychologists view the informational and emotional components of social referencing as separate processes. Specifically, they distinguish between
instrumental social referencing, which involves the learners use of knowledge from others
as indicators of how to understand stimuli (Feinman, 1982), and affective social referencing, which involves the learners use of others emotional facial expressions to determine how to feel about ambiguous events (Klinnert, Campos, Sorce, Emde, & Svejda,
1983).
Learning theorists such as myself, however, have argued that both types of social
referencing are outcomes of the same conditioning process, because the cues that convey
affective components also contain instrumental information, and vice versa. The study
by Gewirtz and Pelez-Nogueras (1992) provided some empirical support for this view,
as well as examples of how the emotional aspects of social referencing can be explicitly
trained.
In contexts of ambiguity, we identified two originally meaningless maternal facial
expressions and then trained them with standard conditioning procedures to denote
opposite consequences for responses where infants reached for objects. Hence, one maternal hand-to-face expression was trained to predict positive auditory-kinetic consequences
of the infant reaching for ambiguous objects (see figure 4.3), while the other maternal
hand-to-face expression was trained to predict negative auditory-kinetic consequences of
the infant reaching for ambiguous objects (see figure 4.4).

the incorporation of a technology based on derived stimulus relations into learning curricula for individuals with developmental disabilities would seem to hold great promise in
helping such individuals acquire functional and meaningful goals.
The present book serves as a compilation of instructional strategies based on decades
of basic and applied research on derived stimulus relations from prominent, worldrenowned researchers who attest to different theoretical frameworks. The book is intended
for parents and a variety of professionals working with individuals with autism and other
developmental disabilities. These professionals include but are not limited to teachers,
developmental therapists, adult service providers, speech-language pathologists, and
behavior analysts, all of whom have some basic understanding of the principles of applied
behavior analysis.
The book is divided into three parts. Part 1 focuses on the establishment of prerequisite
skills necessary for individuals to participate meaningfully in a curriculum based upon,
or including components of, derived relational responding. Part 2 emphasizes instruction
that will lead to the production of such intermediate skills as naming, reading, spelling,
and requesting. Part 3 aims to help the practitioner establish more complex skills in learners, including perspective taking and empathy, higher-order intelligence, and mathematical competence. Each chapter contains a variety of practitioner tools, such as sample data
sheets, step-by-step instructions, training notes, and problem-solving strategies. The reader
need not work through the entire book for it to be of value. Some learners may be more
appropriate candidates for the strategies and techniques presented in one or more parts of
the book only. Thus, the chapters can be used in isolation or in combination with other
chapters, depending on the particular learners educational needs. It is also not necessary
for the reader to be committed to one particular theory regarding derived stimulus relations or verbal behavior, since the chapters represent an eclectic mix of theoretical orientations. Rather, our intention is that the strategies in this book can be incorporated, if not
made the basis of, educational curricula for learners with mild or significant communication and intellectual deficits due to autism, mental retardation, or other developmental
disabilities. As a result, we hope that practitioners and their clients will benefit from the
material presented in this book, and that future years will see the implementation of this
technology in schools, clinics, and habilitation settings around the world.
We wish to thank Anna Neises and Char Burrell for editorial assistance, and we
gratefully acknowledge the many contributors to this book for their thoughtful and persistent work. Many of the authors who contributed to this book are our long-standing
colleagues, friends, and mentors and we have been proud to be associated with their
creativity and expertise. In this vein, we also acknowledge the many researchers, research
assistants, students, and research participants whose work over many years has been an
inspiration for this book.
Ruth Anne Rehfeldt, Ph.D., BCBA
Yvonne Barnes-Holmes, Ph.D.
Southern Illinois University
National University of Ireland,
Carbondale, Illinois Maynooth Ireland

References
Cowley, B. J., Green, G., & Braunling-McMorrow, D. (1992). Using stimulus equivalence procedures to teach name-face matching to adults with brain injuries. Journal of
Applied Behavior Analysis, 25, 461475.
Hanna, E. S., de Souza, D. G., de Rose, J. C., & Fonseca, M. (2004). Effects of delayed
constructed-response identity matching on spelling of dictated words. Journal of
Applied Behavior Analysis, 37, 223228.
Lynch, D. C., & Cuvo, A. J. (1995). Stimulus equivalence instruction of fraction-decimal
relations. Journal of Applied Behavior Analysis, 28, 115126.
Rosales, R., & Rehfeldt, R. A. (2007). Contriving transitive conditioned establishing
operations to establish derived manding skills in adults with severe developmental
disabilities. Journal of Applied Behavior Analysis, 40, 105121.
Sidman, M. (1971). Reading and auditory-visual equivalences. Journal of Speech and
Hearing Research, 14, 513.
. (1977). Teaching some basic prerequisites for reading. In P. Mittler (Ed.), Research
to practice in mental retardation: Vol. 2. Education and training. Baltimore, MD:
University Park Press.
Sidman, M., & Cresson, O. (1973). Reading and crossmodal transfer of stimulus equivalences in severe retardation. American Journal of Mental Deficiency, 77, 515523.

PART 1
Establishing the Prerequisites
for Normal Language

face lights up with surprise and pleasure. However, her next action is not to engage the
kitten in play, but to look up at her mothers face while pointing to the kitten, to see if her
mother had also witnessed the animals dramatic entrance. Gaze shifts may subsequently
be combined with gestures toward the object within the visual field of the familiar face.
Behavioral researchers (including myself) have proposed that an operant (rather than
age-based) history guides the emergence of the skills of joint attention (Dube, MacDonald,
Mansfield, Holcomb, & Ahearn, 2004; Holth, 2005). Specifically, these behaviors normally result from environmental contingencies that operate during early mother-child
verbal and gestural communications (Pelez, Gewirtz, & Wong, 2007). From this perspective, gaze shifts in joint attention incorporate (1) the selective effects of environmental
stimuli that set the occasion for the response class, (2) stimuli that support joint attention
behavioral chains in dual roles as discriminative and reinforcing stimuli, (3) the consequences that lead to the choice of experiencing a stimulus together with the adult versus
experiencing it independently of the adult, and (4) relevant and plausible environmental
conditioning histories. The analysis also identifies the function of reinforcers and suggests
various classes of socially mediated stimuli that maintain joint attention behavior. Indeed,
the most common function of the reinforcers appears to be face-to face interactions with
an adult (Pelez-Nogueras, Field, Hossain, & Pickens, 1996). Put simply, reinforcers are
initially produced by the activity related to the stimulus in question (for example, playing
with a toy) and then increased by adult-generalized social reinforcers such as vocalizations
and smiling, gestures of approval, or demonstrations of affection while engaged. In other
words, it is often more reinforcing for a child to play with a toy or look at a book when
the caregiver participates in the event than it is when the caregiver is absent.

Joint Attention Deficits in Autism

Interest in the concept of joint attention has increased because of its putative role in
developmental disabilities (Carpenter, Pennington, & Rogers, 2002), and Dawson and
colleagues (2004) have even argued that joint attention deficits alone can differentiate
between normally developing learners and those with autism. Specifically, learners with
autism appear to lack prerequisites for joint attention that include orienting to speech
sounds and other social stimuli (for example, when someone points) and show more
direct evidence of deficiencies in joint attention behaviors. For example, Charman and
colleagues (1997) demonstrated that children with autism looked at a mechanical toy
when it was activated but did not exhibit gaze switches between the toy and an adult who
was present.
Deficits in joint attention have also been associated with abnormalities in language
development (Mundy, Sigman, Ungerer, & Sherman, 1986). Specifically, in children with
autism, correlations have been recorded between low frequencies of adult-object gaze
switching at twenty months, limited language gains, and diminished social communication at forty-two months (Charman et al., 1997). One explanation for the relationship
between language and joint attention suggests that the rapid vocabulary expansion of
typical preschool development depends in part on the learners ability to determine, via
observation of adult-attending stimuli, which object in the immediate environment is
related to the adults speech.

According to this view, joint attention should warrant a potentially important place
in early intervention programs, especially those aimed at establishing critical language
prerequisites. And yet it is not often given such an important place in these programs.
Indeed, despite its pivotal developmental significance, there are few effective interventions
for ameliorating deficits in joint attention in the literature. The section below sets out a
training sequence for this purpose that may be used with very young learners or those
with developmental delay.

A Protocol for Establishing Joint Attention

The section below contains a description of the key components of a protocol for
establishing joint attention, followed by empirical evidence to support the use of training
regimes such as these in children with autism.

Establishing social reinforcers. The training of joint attention critically requires the
existence of social stimuli, such as nods or smiles, as reinforcers at an early age. This can
be accomplished with the teacher and learner sitting face-to-face, with ten edible reinforcers spread across the table between them. Block any attempts to remove the reinforcers
from the table until the learner is sitting quietly; then nod and smile before allowing the
learner to take one. It is important to emphasize that the learner is only allowed to take a
reinforcer when the teacher nods and/or smiles (in order to make these gestures function
as discriminative stimuli). In addition, you should emit an occasional verbal cue, such as
yes or Look at that to further improve the learners general communication skills.
Of course, this type of training may lead to the possibility that nods and smiles function as conditioned reinforcers only when treats are available but fail to do so in other
situations. Naturally, this would mean that the learner may not recognize the nods and
smiles of other adults in other contexts. But this situation seems unlikely, or at least relatively easily rectified.
Gaze following. When teaching a child the skill of gaze following, the teacher and
learner should again sit at opposite ends of a small table. First, show the learner a reinforcer of choice, and then ask her to turn around while you place the reinforcer under one
of two opaque cups. Then say, ready, and allow her to turn around again to observe the
cups. Ask her to point to the cup that she thinks contains the treat. Lift the chosen cup,
and if the treat is there the learner can have it. If the empty cup has been selected, simply
remove the treat and start again.
On a subsequent trial, place your face close to the cup with the treat while maintaining eye contact with the learner, such that she comes to rely on this cue for discriminating
the cup that holds the treat. Continue with this type of training until the learner looks
at your face and consistently chooses the cup with the treat. Next, across trials, fade out
your proximity to the cup, so that eventually the learner can choose the right cup after
only a brief glance on your part.
Learners such as those with developmental disorders may experience difficulty simply
attending to others faces. In this case, getting the learner to attend to your face, even
when it is near the cup, will be difficult. In such a situation, it is possible to establish this
skill by saying the learners name, holding the treat up to your eyes, and then tracing

shaping the touch-screen response by modeling, prompting, and signaling appropriate

responding. A familiar assistant may perform these functions for a learner who is severely
disabled. Three stimuli that appear on the screen are directly trained: a picture of an
apple (A), a picture of a baby (D), and a sketch of a bear in a box (G). Each stimulus is
presented in a separate block of trials.

Stage 2: Fading. During this phase of training, comparison stimuli should be gradually faded in, so that each trial begins to more closely resemble the MTS format. That is,
while A, D, or G appear, two comparison stimuli also gradually appear (fade in) until all
three stimuli are clearly visible and the learner can select or point to the target stimulus.
Pointing is reinforced by contingent stimulation (movement and sounds coming from the
stimulus, while the mother or assistant also touches and praises the learner).
Stage 3: Discrimination training. During this stage, all stimuli should appear simultaneously on the screen and the learner must select the appropriate sample. Again, explicit
training of A, D, and G remain in separate blocks.
Stage 4: Identity matching. During identity matching (also called reflexivity training),
A, D, or G each appear individually as a sample with all three stimuli presented as comparisons. The learner is required to select the comparison that is an identity match with
the sample (for example, A-A). Again, each target sample appears within a separate block
of trials.
Stage 5: Mixed identity matching. During this stage, the sample stimuli are randomly
presented within one block of trials. A novel B stimulus (for example, the word apple)
is also introduced as an alternative comparison.
Stage 6: Generalized identity matching. This is a testing stage with no training. A
series of novel stimuli (B, C, E, F, H, and I) appear as random samples within a block of
twelve trials. Accurate identity matching of these stimuli (for example, B-B and H-H) is
deemed evidence of generalized identity matching because none of these had previously
been included during explicit reflexivity training.

of ten) but then became erratic with the introduction of each new stimulus. However,
he did eventually produce seven out of ten correct responses again. Training in identity
matching was extensive but comprised relatively good performances for at least five blocks
of trials. Nonetheless, performance toward the end declined, again suggesting habituation
Response Discrimination
Training Training

Identity
Matching

Mixed
Generalized
Identity Identity
Matching Matching

Reinforcer Identification Strategies and

Teaching Learner Readiness Skills

10
9
8

No. of
Correct
Responses

7
6
5
4
3
Stimulus A
Stimulus D
Stimulus G
Stimuli A, D, G

2
1

Thomas S. Higbee,
Utah State University

Each Block of 10 Trials

and in this case a possible decline in reinforcer efficacy.

Figure 4.2. The data recorded with one learner at each stage of the protocol used by Pelez
et al. (2001).
In summary, training conditional discriminations and training identity matching are
critical precursors to language development, but they are difficult, and there is simply no
easy way to make this happen. The protocol above has been used with some success with
very young or very disabled learners, and common problems encountered in this context
have been noted. The key is to ensure that the target skills can be generalized to novel
stimuli, a capability that will be essential if the learner is to make the crucial transition
between conditional discriminations and derived relations based upon them.

Sample Training Results

In the research reported by Pelez et al. (2001), one child (a twenty-two-month-old
male) readily reached criterion in the initial response training (thirty out of thirty correct
responses) but failed to proceed through discrimination training. Specifically, across nine
blocks of trials, his performance systematically deteriorated from seven out of ten to two
out of ten, at which point the child was removed from the study. In this case, it was
clear that difficulties resulted primarily from fatigue after a training session of longer
than thirty minutes. Notably, this infant demonstrated a significant decline in responding during discrimination training with the introduction of each new stimulus, thus also
suggesting possible habituation.
For illustrative purposes, the data from a second infant (eighteen months old) are
presented in figure 4.2. This child required more extensive response training (sixty trials)
to reach criterion. His performances during discrimination training began well (seven out

CHAPTER 1

The Concept of Joint Attention

The current chapter offers a conceptual and functional distinction between joint attention and social referencing, with the view that the former is a necessary prerequisite for
the development of the latter. Joint attention describes the capacity to use eye contact
and cues to coordinate attention with another person in the sharing of an experience
(such as an interesting object or event; Mundy, Sigman, & Kasari, 1994). Put simply, it
comprises shared awareness of a stimulus. Joint attention begins to emerge between nine
and twelve months of age and initially comprises of gaze shifts between a target object
and a familiar person (Bakeman & Adamson, 1984). Consider a three-year-old girl and
her mother visiting family friends. As the adults sit in the living room and chat, the girl
plays with a puzzle on the floor. Suddenly, a kitten runs into the room and the little girls

Behavioral intervention programs provide students with autism and other developmental
disabilities with opportunities to practice and acquire important skills. The success or
failure of these programs often depends on the quality of reinforcement that is provided
for appropriate student behavior. Though the goal of behavioral intervention programs,
including those based upon derived stimulus relations, is to teach students complex verbal
and social skills, necessary learner readiness skills, such as sitting in a chair and attending
to the instructor and instructional materials, often need to be taught first. This chapter
focuses on strategies for identifying reinforcers and methods for teaching learner readiness
skills.

Reinforcer Identification Strategies

Positive reinforcement is the engine that drives any behavioral education program. For
practitioners working with students with autism, identifying effective reinforcers is often
both the most challenging and the most important part of the intervention program.
Over the past several years, a behavioral technology called stimulus preference assessment
(SPA) has been developed that allows practitioners to identify potentially effective reinforcers for students with autism and other disabilities. Modern SPA techniques involve
systematically providing students with choices between potential reinforcing items and
activities and then measuring their choices. Although multiple methods of conducting

preference assessments may be effective, the two that are most commonly used will be
discussed here: the paired stimulus (forced choice) method and the multiple stimulus
without replacement method.

Paired Stimulus (Forced Choice) Method

The paired stimulus method of preference assessment was developed by Fisher and
colleagues (1992). In the paired stimulus method, potentially reinforcing items and activities are presented two at a time and the student is asked to make a choice between the
two available items. Each item is presented at least twice (once on the left and once on the
right, to control for side preferences) with every other item. A selection response is scored
when the individual makes physical contact with one of the available items (although other
choice responses could be identified for individuals with motor difficulties). Following all
choice trials, a selection percentage is calculated for each item by dividing the number of
times an item was selected by the number of times that it was available for selection and
then multiplying by one hundred (Fisher et al., 1992). Items are then ranked according to
selection percentages. Items with selection percentages of 80 percent or greater are likely
to function as reinforcers.
Fisher and his colleagues (1992) compared the paired stimulus method of assessing
preference to a method called the single item presentation method (Pace, Ivancic, Edwards,
Iwata, & Page, 1985), in which items or activities were presented one at a time to participants, and whether or not the participant made contact with or approached each item
was recorded. Although the single item presentation method had been demonstrated to be
effective in classifying items as being either preferred or not preferred, it was not capable
of directly comparing items to one another (because items were never presented together)
and because of this was thought to overestimate preference in at least some cases. Fisher
and others (1992) conducted preference assessments using both the single item and paired
stimulus methods with the same individuals using the same set of items or activities
for both assessments. They then evaluated the reinforcing effects of items that had been
ranked as highly preferred using both preference assessment methods as well as items that
were ranked as highly preferred by the single item method but not by the paired stimulus
method. The paired stimulus method was shown to more accurately identify potential
reinforcers when compared to the single item method.
The paired stimulus method is generally considered to be the gold standard method
of identifying potential reinforcers. It allows for items or activities to be directly compared
to one another in a controlled and systematic way in order to determine preference. The
only potential disadvantage to the paired stimulus approach is the length of time that can
be required to complete the preference assessment. Pairing each item with every single
other item for two trials can be a very time-consuming process even if only a few items
are compared. For example, twenty preference assessment trials would be required in
order to complete a paired stimulus preference assessment involving five items or activities. If more items are compared, the number of trials goes up exponentially. Thus, even
though it is likely the most accurate form of preference assessment, it may not be practical to conduct paired stimulus preference assessments during each instructional session or
even on a daily basis. Nevertheless, if accuracy is the most important consideration, then

the paired stimulus preference assessment is the most appropriate choice. It may also be
most appropriate for assessing larger numbers of stimuli than can be concurrently assessed
using the MSWO method described below.
The paired stimulus preference assessment data sheet shown below, which is for five
items, can be used to record and analyze the data from the assessment.
The following are the procedures for running the paired stimulus preference assessment (Fisher et al., 1992):
1. Identify four items or activities that the student has requested in the past or has
been observed to interact with during free-choice times, as well as one new item
or activity. If edibles are among the choices, break them up into small bite-size
pieces before presenting them to the student. If a drink is offered present only a
small amount in the cup so the student can drink this amount quickly. If toys
or activities are used, make sure that interacting with the toys or activities will
be meaningful if it only occurs briefly.
2. Allow the student to briefly sample each item by allowing him or her to eat
or drink a small portion of edibles or briefly (for example, for ten to fifteen
seconds) engage with nonedibles.
3. For each trial, place two items on the table about one foot apart in front of the
student. The student should be seated in front of the table with easy access to
the pair. Present each item twice with every other item.
4. Tell the student to pick one. Allow student to approach or select one item.
Circle the item number of the selected item (the first item touched) on the data
sheet (for example, circle 2 if item 2 was chosen over item 3). If the student
simultaneously approaches both items, block access to both items.
5. Remove the nonselected item from the students reach. Allow the student to
consume or have thirty seconds of interaction with the selected item.
6. If the student does not approach or select either item within five seconds,
verbally prompt the student to sample each item. After the sampling period,
present the two items again as directed in steps 2 through 4. If the student still
does not approach or select either item, remove both items, circle N on the data
sheet, and begin the next trial.

MTS training programs in which I have been involved, it is not uncommon to find that a learner appears to fail in one stage but then recovers
or responds well in the next. Does that mean that appropriate responding
in the first stage is present? Also, the prompting and cuing involved are
usually more effective when the mother or teaching assistant is present,
compared to when the learner is working with an unfamiliar adult. In
light of these issues, one minimal benefit to be obtained from the current
protocol is that it offers easy start-up training and confidence building
with young learners who have some limited communicative and instructional histories.
 Ensure that training does not interrupt a regular episode involving sleeping, eating, or changing, because learner attention is frequently influenced by organismic conditions (such as fatigue or hunger).

does not work well with infants, although it usually works well with
learners two years of age and older (see Augustson & Dougher, 1992).
 Reinforcement is provided contingently and immediately on all correct
responses.
 Trials in which the learner makes an incorrect response are not
reinforced.
 Additional interventions (such as shaping and task analyses) are usually
necessary when learners fail to reach criterion after three consecutive
blocks of the same type of trial (see also chapter 8 in this volume).
 Each training block is generally followed by a single block of five randomized probe trials (without feedback).

 Try to make training sessions as short as possible. Fifteen minutes is the

maximum for infants or severely disabled learners if you are to avoid
distraction and fatigue. In my experience, sessions longer than thirty
minutes can contribute to drastic changes and decline in responding.

 Prompts can be used throughout all training trials. However, it is important to determine that the reinforcers, rather than the prompts, are controlling the learners responses, so you must conduct subsequent training
trials with no prompts, and responding should remain the same.

 It is wise to test all reinforcers prior to training. There are numerous standard and simple procedures for doing this (see chapter 1 of this volume).
One type of reinforcer available for automated procedures that has been
found to be useful is the blinking of target stimuli and an accompanying
musical sound (for example, for three seconds).

 Another important aspect to consider is whether the learner can name

the target stimuli. When such a repertoire is not in place, more training trials are usually required (at least with MTS procedures). It may be
useful (though not always an essential or required condition) to teach
stimulus names prior to further training. Naming the objects may facilitate the matching (recognition of the object). See chapter 7 in this volume
for further discussion and instructional strategies for establishing naming
repertoires.

 Try to ensure that interventions are maximally effective by further

enhancing the value of the chosen reinforcers to prevent satiation or
habituation. Two ways to do this include alternating with other reinforcers (Higbee & Pelez-Nogueras, 1998) and ensuring that all reinforcers
are short in duration. Social stimuli such as the caregivers touch, smiles,
and verbal praise have been shown to be very effective with young infants
(Gewirtz & Pelez-Nogueras, 2000).

Training Notes: Presenting Trials

 Learners who fail to respond correctly on all five probes must start training again from the beginning.
 There are some reasons to believe that, for some learners, it may be more
effective to conduct identity matching prior to (rather than after) discrimination training.

Training Notes: Identity Matching

1. Record the number of times each item was selected by totaling the number of
circles in each column of the data sheet.

 Training is usually presented in blocks of ten trials. The mastery-training criterion typically consists of eight consecutive correct responses.
However, with certain learners this criterion may still be too high, and
it is feasible to adjust the criterion (at least early on) to seven consecutive
correct responses. But remember that this is only two responses above
half of all responses being wrong, so try to move back up to the higher
(more stringent) criterion as soon as you can. Do not proceed to the next
stage until the learner has attained this criterion.

The sequence of stages described below are taken from the research by Pelez and
colleagues (2001) and should facilitate the training and testing of conditional discriminations in infants or other persons who have not yet fully developed language. The step-bystep protocol mentions procedural problems that may be encountered and a sample data
set. The six basic stages guide instruction from simple touch-screen response training to
generalized identity matching.

2. Divide the number of times each item was selected by the number of times it
was presented in the SPA (each item is available eight times in a five-item assessment) and multiply by 100 to get a percentage.

 Where there are several comparison stimuli, their locations should be

randomized across trials within a block. However, presenting more than
two comparison stimuli in the teaching of conditional discriminations

Stage 1: Touch-screen response training. This phase of training simply involves teaching the learner to touch a computer screen when a stimulus appears. Young children,
learn faster when seated on their mothers lap. This is because the caregiver is helpful in

7. Repeat steps 3 through 6 until each item has been paired twice with every other
item.
The following are the instructions for scoring the paired stimulus preference
assessment:

cognitive and social processes is clearly reflected in the overlap between learners abilities to
form conditional discriminations, derive relations, develop language, and interact socially
with others. Numerous studies offer empirical support for integration of these core abilities. For example, Devany, Hayes, and Nelson (1986) demonstrated a correlation between
language and equivalence when only the children in their research with no verbal skills
failed to derive equivalence relations. Furthermore, the more severely language-disabled
children also required more extensive training of the target conditional discriminations
than the other children did, thus suggesting that prerequisite abilities in this regard were
also deficient.
A subsequent replication of the study by Devany and colleagues (1986) provided
further evidence of the importance of conditional discrimination abilities to equivalence
and language. In research by Pelez, Gewirtz, Sanchez, and Mahabir (2000), nine normally developing infants, aged twenty-one to twenty-five months, were assessed on the
Receptive-Expressive Emergent Language Scale (REEL-2) and then exposed to a series
of visual-visual conditional discriminations. These involved matching animal-like figures
presented in a match-to-sample (MTS) training format. There were four conditional discriminations: if A then B; if A then C; if D then E; and if D then F. Hence the trained
relations were A-B, A-C, D-E, and D-F. All of the children readily demonstrated the target
conditional discriminations and eight of the children demonstrated transitivity (B-C and
E-F); however, five performed below chance on the symmetry tests (for example, B-A
and F-D). As expected, there was a significant negative correlation between the number
of conditional discrimination training trials and the learners language quotient (in other
words, higher language means less training). These findings highlighted the relationship
between the level of explicit conditional discrimination training necessary for class formation and language competence, and they suggested some degree of distinction between
the various component skills in equivalence.

A Protocol for Establishing Conditional Discriminations

In the relevant literature, there are few studies that have described the explicit training of conditional discriminations in very young learners. This most likely stems from
difficulties in adapting existing methodological paradigms to this population, rather than
weaknesses in the underlying concepts. Indeed, developing experimental methodologies
for use with infants is always extremely challenging, but where they are available they
may offer useful methodologies for training populations with severe developmental disabilities or delays. In one of the only existing studies, Pelez, Lubin, McIlvane, and Dube
(2001) attempted to train and test conditional discriminations in infants who had not yet
fully developed language. One child of eighteen months and two children of twenty-two
months participated. The step-by-step protocol that comprised the training and testing
of these skills is presented below, along with the procedural problems encountered and a
sample data set. The protocol comprises six basic stages that guide instruction from simple
touch-screen response training to generalized identity matching. These, along with an
example of one trial from each stage, are presented in figure 4.1.

3. Rank the items based on the percentages, putting the largest percentage first,
the next largest second, and so forth.

Paired Stimulus Preference Assessment Data Sheet

Illustration of the 6 training phases:

G

APP

BAB

E
Student: Assessed by: Date: Time

Touch-Screen Response Training

Fading

Stimulus items:
1.
2.
3.
4.
5.

Discrimination

APP

APP
APP
BAB

Identity Matching

Mixed Identity Matching Generalized Identity Matching

Figure 4.1. The discrimination training protocol reported by Pelez et al. (2001), with an
example of each of the six stages.

Training Notes: Preparing to Start

Below are some notes and preparations for conditional discriminations training.
 Before you even begin, think very clearly about conducting training of
this sort with very young or very developmentally disabled learners. On
a good day, it can be frustrating. It is extremely difficult to adapt even
the most articulate and systematic methodology to the attention spans of
individuals from these categories, particularly with regard to appropriate responding and attaining a meaningful accuracy criterion. Also, it is
very difficult to preserve their participation and that of their parents or
assistants in a single-learner training environment. These are not small
concerns. These matters raise important questions about the feasibility of
any type of formal training with these learners. They also raise questions
about how to interpret their performances, especially where the methodologies have been adapted from basic laboratory studies. For example, in

Overall rank
(List largest percentage first):

Record item with corresponding item number:

1.
2.
5.
4.
3.
1.
2.
4.
4.
5.
3.
2.
1.
5.
3.
4.
5.
1.
1.
4.
2.
3.
3.
5.
4.
2.
5.
2.
4.
3.
2.
5.
1.
3.
4.
1.
5.
3.
2.
1.
Times selected:
1. x 100 = %
2. x 100 = %
3. x 100 = %
4. x 100 = %
5. x 100 = %

4. Items selected during 80 percent or more of trials are most likely to function as
reinforcers.

Multiple Stimulus Without Replacement Method

Circle item selected:

1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

In an attempt to reduce the amount of time required for preference assessments,

researchers have investigated the effects of offering choices between multiple items or
activities at the same time rather than in pairs. One way to present multiple items has been
termed the multiple stimulus without replacement (MSWO) method. This method was first
developed by DeLeon and Iwata (1996) and later streamlined by Carr, Nicholson, and
Higbee (2000). In a MSWO assessment, multiple (usually five through eight) items or
activities are presented simultaneously. The items or activities are typically placed, equally
spaced, in a row (often called a stimulus array) in front of the student. The student is then
allowed to choose between the items or activities. The item touched first is scored as the
selection and the individual is allowed to consume or interact with the item or activity
for a brief period of time. The selected item is then removed from the stimulus array
and the remaining items are resequenced and the individual is allowed to make another
selection. This process continues until all items have been selected or no item is selected
within a brief period of time (usually five to ten seconds). Usually, this entire process is
repeated three (Carr et al., 2000) to five (DeLeon & Iwata, 1996) times, although comparable results may be obtained in some cases with only completing the selection process
once (Carr et al., 2000). A selection percentage is calculated for each item or activity in
the same fashion as the paired stimulus method, where the number of times an item or
activity is selected is divided by the number of times an item or activity was available
for selection and then multiplied by one hundred. Items are then ranked according to
selection percentage. In contrast to selection percentages obtained from paired stimulus
preference assessments, which can be used to predict reinforcing effectiveness (because the
denominator of the ratio generally remains constant for all items or activities), selection
percentages obtained from MSWO assessments cannot be used to predict reinforcing
effectiveness because the order in which item selections take place can influence selection percentages. For example, if the same item or activity was selected first during each
of three stimulus arrays, its selection percentage would be 100 percent (selected three
times available three times = 100%). In this example, if another item or activity was
selected second during each of three stimulus arrays, its selection percentage would be
50 percent (selected three times available six times = 50%). Thus, selection percentages in the MSWO procedure are used only for ranking stimuli. Items ranked first in
MSWO preference assessments are most likely to function as reinforcers (Carr et al.,
2000; Higbee, Carr, & Harrison, 2000). Data obtained by Higbee, Carr, and Harrison
(1999) also suggest that items ranked second and third also may function as reinforcers
in many cases.
DeLeon and Iwata (1996) compared the MSWO assessment to the paired stimulus procedures developed by Fisher and colleagues (1992). They conducted both types

of preference assessments using the same set of items or activities and with the same
participants. In general, they found that the MSWO preference assessment produced
results comparable to those of the paired stimulus assessment but in about half the time.
Carr et al. (2000) attempted to reduce the amount of time required to complete the
MSWO assessment by reducing the number of stimulus arrays from five to two. They
conducted these brief MSWO procedures with three students with autism and then
examined the reinforcing effectiveness of items or activities identified as being high-,
medium-, and low- preference by the brief MSWO, by delivering these items or activities
contingent on student academic behavior. They found that the brief MSWO procedure
accurately predicted reinforcer effectiveness as contingent delivery of high-, medium-, and
low-preference reinforcers produced responding that corresponded to the degree of preference. In a secondary analysis, Carr and colleagues (2000) calculated correlation coefficients for the stimulus rankings produced by student selections in the first stimulus array
compared to the rankings produced by the combined results of the three arrays and found
that the correlations were high, indicating that conducting an MSWO preference assessment with one stimulus array may be sufficient to accurately rank items or activities. The
authors reported that the brief MSWO assessments could be completed in ten minutes
or less when three stimulus arrays were used. The time could be further decreased if only
one stimulus array was used.
The brief MSWO assessment data sheet shown on the next page can be used to
record and analyze the data from the assessment.
Below are guidelines for conducting the brief MSWO preference assessment (Carr et
al., 2000):
1. Identify four items or activities that the student has requested in the past or
has been observed to interact with during free-choice times, as well as one new
item or activity. If edibles are in the array, break them up into small bite-size
pieces before presenting them to the student. If a drink is offered, present only
a small amount in the cup so the student can drink this amount quickly. If toys
or activities are used, make sure that interacting with them will be meaningful
even if it only occurs briefly.

Brief MSWO Preference Assessment Data Sheet

Pelez-Nogueras, M., Gewirtz, J., & Markham, M. (1996). Infant vocalizations are
conditioned both by maternal imitation and motherese speech. Infant Behavior and
Development, 19, 670.

Student: Assessed by: Date: Time

Pereira-Delgado, J., Greer, R. D., & Speckman-Collins, J. (2006). The effects of using a
mirror to induce generalized imitation. Paper presented as part of a symposium at the
thirty-second annual convention of the Association for Applied Behavior Analysis
International, Atlanta, GA.

Rank by Trial
Stimulus Items

1 2 3 Sum of 1, 2, and 3

Overall Rank (list

smallest sum first)

Student: Assessed by: Date: Time

Rank by Trial
Stimulus Items

1 2 3 Sum of 1, 2, and 3

Overall Rank (list

smallest sum first)

2. Allow the student to briefly sample each item by allowing him or her to eat
or drink a small portion of edibles or briefly (for example, for ten to fifteen
seconds) engage with nonedibles.

Student: Assessed by: Date: Time

3. Place the items on the table or desk in front of the student with equal distance
between them.

Stimulus Items

1 2 3 Sum of 1, 2, and 3

Overall Rank (list

smallest sum first)

4. Provide a brief instruction, such as Pick the one you w ant, to the student, and
allow him or her to choose one item. If the student attempts to grab more than
one item, block access to the other items. You can do this by either pulling the
table or desk out of the students reach or quickly removing all of the nonchosen
items. Write the number next to the item on the data sheet according to the
order in which it was chosen (for example, write a 1 next to soda if soda was
chosen first).

Rank by Trial

Premack, D. (2004). Is language the key to human intelligence? Science, 303, 318320.
Premack, D., & Premack, A. (2003). Original intelligence: Unlocking the mystery of who we
are. New York: McGraw-Hill.
Rogers, S. J., & Pennington, B. F. (1991). A theoretical approach to the deficits in infantile autism. Development and Psychopathology, 3, 137162.
Rosales-Ruiz, J., & Baer, D. M. (1997). Behavioral cusps: A developmental and pragmatic
concept for behavior analysis. Journal of Applied Behavior Analysis, 30, 533544.

CHAPTER 4
Joint Attention and Social Referencing
in Infancy as Precursors of
Derived Relational Responding
Martha Pelez,
Florida International University

Savage-Rumbaugh, E. S., Rumbaugh, D. M., & Boysen, S. (1978). Science, 201, 6466.
Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts.
Skinner, B. F. (1989). The behavior of the listener. In S. C. Hayes (Ed.), Rule-governed
behavior: Cognition, contingencies, and instructional control. New York: Plenum Press.
Tsai, H., & Greer, R. D. (2006). Conditioned preference for books and faster acquisition
of textual responses by preschool children. Journal of Early and Intensive Behavior
Interventions, 3, 3561.
Wolery, M., Holcombe, A., Billings, S. S., & Vassilaros, M. A. (1993). Effects of simultaneous prompting and instructive feedback. Early Education and Development, 4,
2031.

Joint attention and social referencing appear to be critical features of parent-child or

teacher-learner exchanges, which help the learner gather information to guide his own
thoughts, feelings, and behavior. These two related abilities also appear to be necessary
precursors for the emergence of derived relational responding, which underpins much
of higher cognition and language development. The current chapter focuses on teaching
strategies for establishing these core skills with very young learners or learners with developmental delay, including those with autism spectrum disorder. The first section of the
chapter discusses strategies, including a protocol, for the establishment of the prerequisite
skills for joint attention and social referencing in terms of training conditional discriminations and identity matching. The second section describes behavioral interventions and
strategies for establishing the core features of joint attention and social referencing.

Establishing the Prerequisite Skills for Joint

Attention and Social Referencing
As well as constituting core social skills, joint attention and social referencing are
codependent abilities that are derived from basic visual discriminations, which have long
been known to be prerequisites for derived relational responding. This early integration of

Collier, C. K., & Bitetti-Capatides, J. (1979). Positive behavioral contrast in three-monthold infants on multiple conjugate reinforcement schedules. Journal of Experimental
Analysis of Behavior, 32, 1527.

symposium at the thirty-second annual convention of the Association for Applied

Behavior Analysis International, Atlanta, GA.

Darcheville, J. C., Madelain, L., Buquet, C. Charlier, J., & Miossec, Y. (1999). Operant
conditioning of the visual smooth pursuit in young infants. Behavioural Processes, 46,
131139.

Greer, R. D., Keohane, D. D., & Delgado, J. P. (2006). Conditioning adult voices to induce
listener capabilities. Paper presented as part of a symposium at the thirty-second
annual convention of the Association for Applied Behavior Analysis International,
Atlanta, GA.

DeCasper, A. J., & Fifer, W. P. (1980). Of human bonding: Newborns prefer their
mothers voices. Science, 208, 11741176.

Greer, R. D., & Ross, D. E. (2008). Verbal behavior analysis: Inducing and expanding new
verbal capabilities in children with language delays. New York: Allyn and Bacon.

Dinsmoor J. A. (1985). The role of observing and attention in establishing stimulus

control. Journal of the Experimental Analysis of Behavior, 43, 365381.

Greer, R. D., Singer-Dudek, J., & Gautreaux, G. (2006). Observational learning.

International Journal of Psychology, 41, 486499.

Donahoe, J. W., & Palmer, D. C. (2004). Learning and complex behavior. Richmond, VA:
Ledgetop Publishing.

Hart, B., & Risley, T. R. (1999). The social world of children learning to talk. Baltimore,
MD: Paul H. Brooks Publishing.

Eimas, P. D., Siqueland, E. R., Jusczyk, P., & Vigorito, J. (1971). Speech perception in
early infancy. Science, 171, 303306.

Hayes, S., Barnes-Holmes, D., & Roche, B. (2001). Relational frame theory: A postSkinnerian account of human language and cognition. New York: Kluwer Academic/
Plenum.

Epstein, R. R., Lanza, P., & Skinner, B. F. (1980). Symbolic communication between two
pigeons (Columbia livia domestica). Science, 207, 543545.
Epstein, R. R., Lanza, P., & Skinner, B. F. (1981). Self-awareness in the pigeon. Science,
212, 695696.
Field, T. (1987). Affective and interactive disturbances in infants. In J. D. Osofsky (Ed.),
Handbook of infant development (2nd ed.). Oxford, England: Wiley.
Garcia, E., Baer, D. M., & Firestone, I. (1971). The development of generalized imitation
within topographically determined boundaries. Journal of Applied Behavior Analysis,
4, 101112.
Greer, R. D. (2002). Designing teaching strategies: An applied behavior analysis system
approach. San Diego, CA: Academic Press.
Greer, R. D., Chavez-Brown, M., Nirgudkar, A. S., Stolfi, L., & Rivera-Valdes, C. L.
(2005). Acquisition of fluent listener responses and the educational advancement of
young children with autism and severe language delays. European Journal of Behavior
Analysis, 6, 88126.
Greer, R. D., Dorow, L. G., & Hanser, S. (1973). Music discrimination training and the
music selection behavior of nursery and primary level children. Bulletin of the Council
for Research in Music Education, 35, 3043.
Greer, R. D., Dorow, L. G., Wachhaus, G., & White, E. (1973). Adult approval and students music selection behavior. Journal of Research in Music Education, 21, 293299.
Greer, R. D., & Keohane, D. D. (2005). The evolution of verbal behavior in children.
Behavior Development Bulletin, 1, 3147.
Greer, R. D., Keohane, D. D., Ackerman, S., OSullivan, D., Park, H., Longano, J., et
al. (2006). Sensory matching protocol: Providing children with the capacity for sameness
across the senses as a component of basic listener literacy. Paper presented as part of a

Horne, P. J., & Erjavec, M. (2007). Do infants show generalized imitation of gestures?
Journal of the Experimental Analysis of Behavior, 87, 6387.
Horne, P. J., & Lowe, C. F. (1996). On the origins of naming and other symbolic behavior. Journal of the Experimental Analysis of Behavior, 65, 185241.
Keohane, D. D., & Greer, R. D. (2005). Teachers use of a verbally governed algorithm
and student learning. International Journal of Behavioral and Consultation Therapy, 1,
252271.
Keohane, D. D., Greer, R. D., & Ackerman, S. A. (2006). The effects of conditioning
visual tracking on the acquisition of instructional objectives by prelisteners and prespeakers. Paper presented as part of a symposium at the thirty-second annual convention of
the Association for Applied Behavior Analysis International, Atlanta, GA.
Krekling, S., Tellevik, J. M., & Nordvik, H. (1989). Tactual learning and cross-modal
transfer of an oddity problem in young children. Journal of Experimental Child
Psychology, 47, 8896.

5. When the student has finished consuming the edible or the specified period
of interaction has elapsed for engagement with the nonedible item or activity,
remove the nonedible item or activity and place it out of sight. Arrange the
remaining four items as in step 3 and center them in front of the student.
6. Steps 3 through 5 will be repeated until all items have been selected and no
items are left, or until the student does not select an item within ten seconds. If
the student fails to select an item within ten seconds, score all of the remaining
items as 5 on the data sheet.
7. Repeat the entire procedure two additional times using the same items (step 2 is
not necessary after the first array). Record data in column 2 for the second array
and in column 3 for the third array. If you find that the student is responding
in similar fashion to all three stimulus arrays, you may be able to shorten the
procedure to one array in subsequent assessments.
The following are the instructions for scoring the the brief MSWO preference
assessment:
1. Add the ranks for each item in columns 1, 2, and 3 and then record this number
in the Sum of 1, 2, and 3 column.
2. Rank the items based upon the numbers in the Sum of 1, 2, and 3 column,
with the smallest number being ranked first, the next smallest being ranked
second, and so on.

Answers to Common Questions About

Preference Assessments
Q: How often should I run the preference assessment?

Kuhl, P. K., & Meltzoff, A. N. (1996). Infant vocalizations in response to speech: Vocal
imitation and developmental change. Journal of the Acoustical Society of America, 100,
24252438.

A: Preferences have been shown to be relatively stable for some students and fluctuate
greatly for others (Carr et al., 2000). A conservative approach would be to conduct preference assessments at least daily. It would be preferable to complete a preference assessment
multiple times per day, such as before each teaching session or when the students performance starts to deteriorate, if possible.

Meltzoff, A. N., & Moore, M. K. (1983). Newborn infants imitate adult facial gestures.
Child Development, 59, 702709.

Q: How do I choose the items to use in the preference assessment? Why should I
include a new item each time?

Mundy, P., Sigman, M., & Kasari, C. (1994). Joint attention, developmental level, and
symptoms presentation in autism. Development and Psycholopathology, 6, 389401.

A: Watching what your student interacts with during free play is a good way to
select items for inclusion in the preference assessment. Informal interviews with parents
or other caregivers can also provide information about what to include in the assessment.
It is important to include new items so that the student is exposed to them during the
stimulus sampling procedure. Continuing to try new items in a search for new potential
reinforcers is also important.

Novak, G., & Pelez, M. (2004). Child and adolescent development: A behavioral systems
approach. Thousand Oaks, CA: Sage Publications.

Q: Is it okay to assess edibles and nonedibles together in the same preference

assessment?
A: Combining edibles and nonedibles in the same preference assessments may be
problematic in some cases, since some students tend to select edible items before nonedible items, even though the nonedible items may actually function as reinforcers (DeLeon,
Iwata, & Roscoe, 1997). Thus, if your student appears to be selecting all of the edible
items before the nonedible items, you may consider conducting separate preference assessments for edibles and nonedibles.
Q: Do I need to use the actual items or activities in the preference assessment? Is it
okay to use symbols or pictures of the items or activities or conduct an entirely verbal
preference assessment?
A: Several studies have been conducted in an attempt to answer these questions.
In these studies, preference assessments have been conducted using pictures of stimulus
items or symbols, either representing these items instead of the items themselves (for
example, presenting pairs or arrays of pictures of potential reinforcers and asking students
to choose which one they would most like to earn as a reward for working) or presenting
potential reinforcers in a verbal forced-choice format (for example, Would you like to
work for candy or music?). Results of research on the use of verbal-, picture-, or symbolbased preference assessments have been mixed, with some studies showing positive effects
(Graff & Gibson, 2003) and others not (Higbee et al., 1999). A critical variable appears
to be whether or not access to the chosen item or activity is provided immediately following a selection response in the preference assessment. Preference assessments appear to be
more accurate when access to the chosen item is provided following a selection response
(Tessing, Napolitano, McAdam, DiCesare, & Axelrod, 2006). In summary, when possible, it is best to use the actual items or activities in preference assessments. Pictures,
symbols, and verbal representations in preference assessments should be used with caution
until further research determines the conditions under which they can most effectively
be used.

Summary: Strategies for Reinforcer Identification

The identification of effective reinforcers is a critical task for practitioners attempting to
provide instruction to students with autism and other developmental disabilities. Finding
reinforcers for children with autism can be particularly difficult due to the restricted
patterns of interest that often accompany the disorder. It is also particularly important,
however, because the most successful educational interventions for these children rely
heavily on procedures based on positive reinforcement. Students in intense behavioral
programs, including those using the procedures described in this book, are presented
with hundreds of learning opportunities per day. Powerful positive reinforcers are necessary to provide motivation for learning and positive consequences for correct responding.
Stimulus preference assessments, like those discussed in the present chapter, are a valuable
tool for helping practitioners identify effective reinforcers for students with autism and
other developmental disabilities.

Teaching Basic Learner Readiness Skills

Now that they have powerful and effective reinforcers in hand, the next task for practitioners working with students with autism and other developmental disabilities and hoping
to ultimately establish derived relational responding skills is to start teaching basic readiness skills. Before tackling such complex skills as language, social, and intellectual skills,
many students will need to learn prerequisite skills that will allow them to be active
participants in the learning process. These learner readiness skills include such behaviors
as sitting appropriately in a chair, attending to the instructor and instructional materials, and imitating simple responses. Variations on two basic behavioral techniques
shaping or differential reinforcement and promptingare typically employed to teach
these important prerequisite skills.

Differential Reinforcement and Prompting Techniques

Differential reinforcement is the act of delivering reinforcement for certain specific
responses while not reinforcing other responses. The most common form of differential
reinforcement that is used in behavioral education programs is called shaping. Shaping is
a specific differential reinforcement procedure where successive approximations to a target
behavior are differentially reinforced. Instead of expecting the student to perform the
final target behavior, which may be beyond the students capabilities at the moment, reinforcement is provided for performing an approximation or intermediate form of the target
behavior. Once the student can reliably produce the approximation of the target behavior,
the reinforcement criterion is changed so that the student must produce a behavior that
more closely resembles the target behavior. In this way, responding is gradually shaped
into forms that more and more closely resemble the target behavior until the ultimate
target behavior is reached. Shaping allows the student to come in contact with the reinforcer for the target response early and often and helps lead the student to the ultimate
target behavior. For example, if we were teaching a student to respond to the simple direction Raise your hand by fully extending his or her arm and hand upward (the target
response), one might initially reinforce any hand movement in response to the verbal
instruction. Once this behavior was occurring reliably, we might then require the student
to raise his or her hand to chest level with the hand open in order to receive reinforcement. Finally, we could then require the complete action of fully extending the arm and
hand in response to the teacher instruction. If the student had difficulty with any of these
shaping steps, we could create smaller approximations.
Prompting is another instructional technique that is used to teach learner readiness
behavior. Prompting involves providing the student with additional help (prompts) beyond
the initial instruction, which promotes correct responding. Prompts come in all shapes
and sizes and fall into two basic categories: response prompts and stimulus prompts
(Cooper, Heron, & Heward, 2007). Response prompts are most common and include providing the student with additional verbal instructions, a model of the correct response,
or physical guidance or assistance. For example, if we were teaching a student to clap his
or her hands in response to the instruction Clap your hands, we could provide a model

prompt by clapping our hands following the instruction (and before the student makes
an error). The student could then imitate the response. Over time, we would fade out, or
gradually eliminate, the model prompt so that the student would respond to the instruction alone. One way to do this would be to use a time-delay procedure in which we
gradually introduced an increasing delay between our instruction and the model prompt
(Halle, Marshall, & Spradlin, 1979).
Stimulus prompts involve highlighting or changing some aspect of the instruction or
instructional stimuli provided to the student, such as using position prompts (where the
correct stimulus is placed closer to the student), or altering the instructional stimuli to
make the correct response more apparent. For example, when teaching a student to select
the appropriate letter from an array of three letter cards in response to the instruction
Give me the letter A, the A on the card could be made larger than the letters on the
other cards, with the size of the letter gradually reduced across instructional trials (but
see Doran & Holland, 1979; Sidman & Stoddard, 1967; Stoddard & Sidman, 1967;
Touchette, 1968). It should be noted that the prompting-and-fading procedure chosen
should be individualized for the particular student. When working with students with
severe disabilities, practitioners may find that one procedure (such as using added brightness or size prompts that are then faded out gradually) is ineffective for a particular student
but effective for another. Moreover, responding may come under exclusive control of the
particular prompt that was added and it may not be clear that this has occurred until
errors occur in the final step(s) of fading when the prompt becomes unavailable. For this
reason, prompting strategies must be chosen with careful attention paid to the students
prior history with prompting systems and idiosyncracies in his or her attention to certain
relevant or irrelevant aspects of stimuli.

during instruction (for example, tap fingers on desk, rub hands together, lift up one leg).
Subsequently, the mirror procedure was implemented to determine if the children would
acquire generalized imitation. The participants were closely matched based on levels of
verbal behavior with peers who did not receive the mirror procedure. The findings paralleled the results of the first experiment; moreover, the matched peers who did not receive
the mirror intervention did not acquire generalized imitation even when the numbers of
learn units were controlled for.
The results of both experiments can be explained in terms of the correspondence
between see and do, a theory suggested by Catania (1998) as something that could
be established using a mirror. It seems plausible that the opportunities the children had
to view their responses in the mirror enhanced this see-and-do correspondence. Prior
research suggests that deficits in imitation may result when an individual is not able to
observe a visible endpoint (Meltzoff & Moore, 1983). For example, a child may have no
difficulty imitating clapping because the response is visible to him. Touching ones head,
however, may be more difficult for him to imitate because the response is out of the
childs sight. The mirror provides children with the opportunity to see their responses. As
noted previously, prior research has found that children with disabilities typically require
an extensive amount of shaping to develop imitation of discrete behavior, yet this does not
guarantee that they will acquire generalized imitation. For example, Baer and colleagues
(1967) found that children with developmental disabilities did not acquire generalized
imitation until their imitative repertoire consisted of at least forty to sixty directly trained
imitative responses! The mirror protocol, outlined in table 3.7, has shown promise in
establishing generalized imitation skills efficiently and expediently.

TABLE 3.7. Generalized Imitation

through the Mirror Protocol

Teaching Appropriate Sitting Behavior

Sitting in a chair is often one of the first skills targeted in behavioral programs for
students with autism and other developmental disabilities. Not only is sitting an important prerequisite behavior, but the procedures used for teaching it, physical prompting and
differential reinforcement, are relatively straightforward. Thus, it is often a quick win
for both the instructor and the student and allows the student to contact positive reinforcement quickly. It is important to note, however, that this is not always the case, and
that the physical prompting involved in teaching students to sit in a chair may cause some
students to engage in disruptive behavior such as aggression, screaming, and noncompliance. Although the specific procedures that will be most effective in decreasing these
negative behaviors will vary for each individual student, the general strategy typically
involves shaping appropriate sitting by providing reinforcement for increasingly longer
periods of time where the student is sitting appropriately without engaging in disruptive
behaviors. What is initially most important is that the student obtains reinforcement
while engaging in appropriate sitting behavior (in other words, the reinforcer is delivered
before the student engages in negative behavior). The initial time intervals may be very
short (a few seconds) but can usually be increased relatively quickly if the reinforcer used
is powerful and the student regularly gains access to it.

Rationale

If the child does not show generalized imitation, use this protocol.

Long-Term Given a set of 20 unreinforced test trials, which consist of novel fine
Objective and/or gross motor actions presented by the instructor, the child will
imitate these actions with 80% accuracy for 1 session.
Materials

A full-length unbreakable mirror measuring approximately 13.5 by 50

inches is used for this protocol. Note: Breakable mirrors may be hazardous to the health of children and professionals.

Special Note For the probe sessions, reserve a set of 20 novel motor actions that are
not in the students repertoire (do not teach these actions in the mirror).
For the teaching sessions with the mirror, if the child has mastered
several directly taught actions prior to the implementation of this protocol, teach these previously mastered actions in the mirror; otherwise, you
can teach novel actions in the mirror (not part of the probes).

Ensure that the child is looking at the mirror and not at the teacher
by pointing to the mirror as part of the antecedent during instructional
sessions. Suspend all other instructional programs on imitation.
General
Procedure

Conduct unconsequated probes (20 novel actions) directly facing the

student (do not use the mirror). Begin with the first short-term objective
of 4 target actions to be taught in the mirror using 20-learn-unit sessions until criterion is achieved. Conduct a post-probe session for novel
actions. Continue with a new short-term objective of 4 novel target
actions to be taught in the mirror until generalized imitation is established or the long-term objective is achieved during the probe sessions.

Criterion

90% correct learn units for two consecutive sessions for teaching actions
in the mirror. 80% correct trials for probes for one session, which is the
LTO for this protocol.

Research to a Theory of Verbal Development

Dinsmoor (1985) noted that greater observing and attending to specific environmental
stimuli resulted in an increase in control over behavior by those stimuli. In this chapter
we have presented what we believe to be strong evidence of the establishment of childrens
observing responses as a result of the implementation of the conditioning procedures and
protocols described. After implementation of these protocols, the children with whom
we have worked have advanced to higher levels of verbal capability. We also believe that
the evidence illustrating the efficacy of these protocols presents a more complete view of
the foundations of verbal behavior. In conclusion, we submit that the role of the listener
as observer and particularly the interrelationships among listener, listener-speaker, and
speaker-as-own-listener functions are the foundations basic to the development of early
language capabilities (Greer & Keohane, 2005).

References
Baer, D. M. (1983). Can you decode a code? Behavioral and Brain Sciences, 1, 138139.
Baer, D. M., & Deguchi, H. (1985). Generalized imitation from a radical-behavioral
viewpoint. In S. Reiss & R. Bootzin (Eds.), Theoretical issues in behavior therapy. New
York: Academic Press.
Baer, D. M., Peterson, R. F., & Sherman, J. A. (1967). The development of imitation by
reinforcing behavioral similarity to a model. Journal of the Experimental Analysis of
Behavior, 10, 405416.
Brigham, T. A., & Sherman, A. J. (1968). An experimental analysis of verbal imitation in
preschool children. Journal of Applied Behavior Analysis, 1, 151158.
Catania, A. C. (1998). Learning (4th ed.). Englewood Cliffs, NJ: Prentice Hall.

Criterion

Given one positive exemplar (an accurate match) and two nonexemplars
(nonmatching stimuli), children will match identical exemplars of the
target items, rotated across gustatory, visual, olfactory, and tactile senses
until they achieve standard mastery criterion (typically 90% across 2
sessions). Once the child has met the long-term objective, you may
return to the full schedule of subject-area lessons. The rate of learning
should have accelerated such that children have significantly reduced the
numbers of learn units required to master instruction.

In the studies we have conducted to date, childrens levels of observing responses

have been shown to increase across the areas targeted when this protocol is implemented
(Greer, Keohane, Ackerman, et al., 2006). After completion of the sensory matching
protocol, children are often able to master new skills, including pointing and echoing.
We also assume that the steps required for the prelistener level of verbal capability have
now been achieved, and the child is ready to begin the next set of protocols as he moves
toward a more comprehensive listener status. (See Greer & Ross, 2008, for additional
information on the sensory matching protocol.) The original focus of this procedure was
simply to develop a capacity for sameness, but we found that the procedure frequently had
collateral effects on observing responses. We do not yet know why this appears to be the
case for some children and not others.

Generalized Imitation as the Next Step

The acquisition of imitation skills is critical for the development of language and social
repertoires (Rogers & Pennington, 1991). Baer, Peterson, and Sherman (1967) defined
imitation as a behavior that closely follows another individuals behavior in which the
form is controlled by the behavior of other individuals. For example, a child who imitates will likely clap his hands when a model claps his hands if the response has been
directly trained. Imitation is a see-and-do relationship, which involves a point-to-point
correspondence between behavior of the model and the behavior of the observer (Greer
& Ross, 2008). Generalized imitation is present when a child imitates novel behavior
without direct reinforcement of that particular response. Generalized imitation, unlike
observational learning, is the result of direct reinforcement of a class of responding and
is not controlled by the observation of the contingencies experienced by another (Greer,
Singer-Dudek, & Gautreaux, 2006). Generalized imitation thus involves the continued
performance of a response for which the child has never received reinforcement (Brigham
& Sherman, 1968). Once a child has acquired generalized imitation, he does not need to
be explicitly taught each motor action separately with response prompts. For example, a
child may sit down, clap his hands, and wave bye-bye once he has learned the behavior
of imitatingeach response does not need to be explicitly shaped. Generalized imitation can thus be regarded as a higher-order operant (Greer & Ross, 2008). A teacher can
simply demonstrate a behavior and then teach the relevant antecedent and consequent
relations that are needed to teach a new operant.

Much research has been devoted to directly taught imitation and generalized imitation in both typically developing children and children with developmental disabilities. Typically developing children at age four have shown generalized imitation skills
(Baer & Deguchi, 1985), but infants of one and two years of age have not (Horne &
Erjavic, 2007). Children with severe developmental delays often have deficits in their imitative repertoires. Research over the past few decades has identified successful tactics to
teach imitative responses to such children. These include simultaneous stimulus prompts
(Wolery, Holcombe, Billings, & Vassilaros, 1993) and shaping procedures (Garcia, Baer,
& Firestone, 1971). These tactics are often successful in teaching an explicitly defined
set of responses. For some children such strategies lead to the development of generalized imitation, but for other children such techniques are often tedious and unsuccessful
because the class of responding or higher-order operant has not been formed.
In our schools we use learn units, as described earlier in this chapter, to measure
teacher and child responses. In many cases learn units combined with the above-mentioned strategies do not result in generalized imitation. For example, we may teach children to imitate the following set of actions to mastery using the learn unit: stand up,
clap hands, touch head, and wave. Even though the children are explicitly taught these
responses, they may not imitate novel actions, such as sit down, tap lap, and touch shoulders, which would be indicative of generalized imitation. Children who do not display
generalized imitation repertoires are considered prelisteners. Imitation and generalized
imitation are two of the five basic attention programs taught in our schools. They are
prerequisites for learning basic listener responding. Until a child acquires listener literacy
(Greer, Chavez-Brown, Nirgudkar, Stolfi, & Rivera-Valdes, 2005), he will not achieve
more-advanced developmental milestones. Listener literacy includes the ability for ones
behavior to be governed by the speaker behavior of others and results in a greater degree
of independence for the child (Greer, 2002).

The Mirror Protocol

Pereira-Delgado, Greer, and Speckman-Collins (2006) evaluated the use of a mirror
in the emergence of generalized imitation. In the first experiment, the participants were
three- and four-year-old children diagnosed with developmental disabilities who had
acquired numerous motor actions through direct learn-unit presentations combined with
various other instructional tactics. However, the children did not show generalized imitation during unreinforced test trials. The intervention consisted of teaching the children to
imitate teacher actions using a mirror, in which the child faced the mirror and the teacher
sat behind the child. Learn-unit presentations were used throughout the mirror procedure, where the child was required to look in the mirror to observe the teachers presentation of a given motor action. The target behavior included the child modeling or copying
the teachers motor actions with a point-to-point correspondence. We found that, after
being taught a set of four previously acquired motor actions in the mirror to mastery,
the participants acquired generalized imitation. In the second experiment, the participants were also three- and four-year-old children with developmental delays. However,
these children were only taught a few motor actions directly. Test sessions for generalized
imitation included a specified set of twenty different motor actions that were not taught

Below are guidelines for teaching sitting:

1. Sitting is usually taught by first placing two chairs facing one another (one for
the student and one for the teacher), one to two feet apart.
2. The student is guided to stand in front of his or her chair, facing the instructor
while the instructor is seated. To prevent tipping, the instructor may place the
students chair with its back against a wall or may use his or her legs and feet to
keep the chair on the ground.
3. The instructor then gives a verbal instruction such as Sit or Sit down and
then immediately provides a physical prompt to assist the student in sitting.
4. When the student sits appropriately, even if prompts were provided, reinforcement is delivered. Reinforcement should be delivered while the student is
sitting.
5. Prompts should be faded as quickly as possible as the student becomes more
independent at sitting. The amount of reinforcement provided should roughly
correspond to the amount of independence the student demonstrates in sitting.
In other words, the less prompting required to get the student to sit, the higher
the quality and quantity of reinforcement that should be provided.
For more detailed information about teaching the skill of sitting, see chapter 9 in O. I.
Lovaass book Teaching Individuals with Developmental Delays (2003).

Teaching Eye Contact and Other Attending Behaviors

The ability to attend to an instructor and instructional materials is another important prerequisite skill for students with autism and other developmental disabilities to
learn. Historically, attending behaviors have been explicitly taught by presenting verbal
cues such as Look at me and then providing differential reinforcement for increasingly
longer durations of eye contact (Lovaas, 1981; Taylor & McDonough, 1996, p. 74). When
students in these studies did not readily provide eye contact, increasing levels of prompting were provided, which included displaying a reinforcer (usually an edible item) and
then pulling the reinforcer up close to the instructors face. When the student provided
eye contact for the specified period of time, the reinforcer was delivered. This procedure
was typically repeated several times in succession.
Although procedures such as this can be effective in teaching students to provide eye
contact in response to a verbal instruction, they seem rather contrived. Most typically
developing three- and four-year-olds would not look at a parent for five to ten seconds in
response to their name or following an instruction such as Look at me. In fact, working
on eye contact in isolation may actually take valuable instructional time that could be
used for teaching other skills. More recent approaches to teaching eye contact and attending involve teaching these skills within the context of other instructional programs. For
example, instead of teaching the student over repeated trials to give eye contact for longer
and longer intervals in response to a verbal instruction (the Look at me approach),
many behavioral interventionists work on eye contact while they are concurrently teaching

imitation, receptive object identification, matching, and other tasks. The prompting procedures used in this contextual approach are often the same as in the Look at me
approach, although prompting procedures are often used instead of directly showing the
student a reinforcer to attract his or her attention. Such prompting procedures include
time delay (waiting for a few seconds for the student to give eye contact), using instructional materials to attract the students attention (placing a flash card or other object in
the students line of sight near his or her face and then slowly moving the object back near
the instructors face), and light physical prompting (lightly touching the students face and
then directing attention to the instructors face by the instructor pulling his or her hand
back toward his or her eyes).
The primary difference is in the consequences provided for giving eye contact. In
the Look at me approach, reinforcement is provided when eye contact is given. In the
contextual approach, when the student gives appropriate eye contact in the instructional
setting, he or she is presented with an opportunity to respond to a direction and then
reinforcement is provided, contingent upon his or her responding correctly to the direction. Thus, eye contact is conceptualized as a necessary part of the instructional sequence
rather than a separate skill to be taught in isolation. In addition to requiring eye contact
before presenting the student with an opportunity to respond to an instruction, practitioners should also provide reinforcement when the student makes eye contact spontaneously
in the course of instruction or play. In many cases, simply requiring students to be either
attending to instructional materials or looking at the instructor before he or she gives an
opportunity to respond is sufficient to produce the necessary amounts of eye contact and
attending behavior.
Another contextually appropriate situation for teaching students to make eye contact is
in teaching students to mand for, or request, preferred edibles, objects, or activities. Mand
training takes advantage of naturally occurring student motivation to request preferred
items or activities (often edibles in early stages of training). The student is prompted to
make an appropriate communicative response via vocal speech, sign, or picture exchange
(depending on the communication level of the student) and then the requested item is provided contingent on the response. Once the student reliably makes requests, the instructor can add an eye contact requirement in addition to the communicative response. Thus,
in order to obtain the requested item, the student must provide eye contact while making
the communicative response. Using differential reinforcement, the instructor can teach
the student to provide eye contact when making requests.
Two important factors can influence student eye contact and attending behaviors:
motivation and pacing of instruction. If the instructor is providing frequent enough access
to powerful reinforcers contingent on correct student responding, the student will likely
be attentive to the instructor and materials because he or she is motivated to gain further
access to the reinforcers. If the instructor is providing reinforcement too infrequently
or if the items or activities being delivered are not potent reinforcers, the instructor will
likely be struggling to keep the students attention. Also, appropriately paced instruction
helps students to maintain attention to the instructor and instructional materials. Though
student attention can be lost when instruction is happening too rapidly, it is most often lost
when the pace of instruction is not rapid enough. Thus, a good strategy for m aintaining
student interest and attention is to be sure to have potent reinforcers (identified using a
preference assessment), deliver these reinforcers on a rich enough schedule, and keep the
pace of instruction high.

Below are guidelines for teaching and maintaining eye contact and attending
behavior:
1. Teach eye contact and attending within the context of other instructional
programs.
2. Before presenting an instruction, require the student to look at you or at the
materials.
3. Start with a time-delay prompt. Wait up to five seconds for the student to
provide eye contact on his or her own.
4. If the student does not readily provide eye contact, use instructional materials
or a light physical prompt to obtain the students attention.
5. When the student provides attention, immediately provide him or her with an
opportunity to respond to a direction.
6. Provide reinforcement when the student responds correctly to the direction
provided.
7. Also provide reinforcement when eye contact is provided spontaneously during
play or instruction.
8. Work on eye contact while teaching mands for preferred objects by requiring
students to provide eye contact when they are making requests.
9. Use high-quality reinforcers and keep the pace of instruction rapid enough.

Teaching Generalized Imitation Skills

Typically developing children learn many simple and complex skills by observing
and imitating the behavior of others. Many students with autism and related disorders,
however, do not readily imitate the behavior of others. Thus, teaching the student to
imitate a model is another important area of learner readiness programming. Because
model prompts are often used in the teaching of more advanced skills, learning to imitate
is critical for the development of more complex behaviors. The goal of imitation training
is to produce what has been called generalized imitation, where the student will reliably
imitate new behaviors without specific training on the modeled response (Lovaas, 2003;
Peterson & Whitehurst, 1971). Imitation skills are often broken down into two categories:
nonverbal (motor) imitation and verbal imitation.
Nonverbal imitation is typically one of the first instructional programs implemented
with young children with autism. The only prerequisite skills required for participation in
nonverbal imitation training are the ability to sit in a chair and attend to the instructor,
although, as discussed previously, these skills can be taught concurrently during imitation
training. As is the case with other instructional programs, nonverbal imitation training
usually starts with simple motor responses (such as putting a block in a bucket or clapping
ones hands) and then progresses to more complex behaviors (such as play and other social

behaviors) as the student acquires basic imitative skills. During initial imitation training,
many students have more success learning to imitate motor actions that involve objects,
such as putting a block in a bucket or hitting the table with a toy hammer, as opposed to
learning to imitate actions that only involve body movements, such as waving good-bye or
nodding the head. When teaching targets that involve the manipulation of objects, however,
it is important to also have distracter objects on the table to ensure that the student is imitating the instructors behavior rather than simply learning that when the block and the
bucket are together on the table, the block goes in the bucket (Lovaas, 1981).
The typical teaching arrangement is similar to others that have been discussed: The
teacher and student are seated and facing one another. Often, there is a table for materials
that is either in between the instructor and the student or to the side. Each instructional
trial begins with the instructor providing a verbal cue such as Do this while simultaneously performing the motor action that the student is being asked to imitate. For
example, if the motor response involves placing a wooden block in a bucket, the instructor
states, Do this, while dropping the block into the bucket. The student is then physically
prompted to imitate the motor response and reinforcement is provided when he or she
does so. As with other teaching procedures discussed previously, the amount and quality
of reinforcement should correspond to the quality of the behavior. Thus, more-independent
responses should produce higher-quality reinforcement. In some cases, it may be necessary
to have a second instructor prompt the student while the primary instructor models the
correct response. Over time, prompts are faded until the student is reliably imitating the
motor response. The amount of target responses being concurrently taught will vary from
student to student and may be as small as one or as high as four or five responses. As target
responses are mastered, new targets are then introduced. When several motor responses
have been learned, untrained items should be presented to test for generalized imitation.
Verbal imitation training teaches the student to imitate a verbal model (such as a
sound, word, or sentence) rather than a motor response. By teaching verbal imitation, the
teacher can bring the students vocal behavior under instructional control. This means
that the student will produce sounds in response to the model provided by the instructor rather than for other reasons, such as self-stimulation. According to Lovaas (2003),
bringing the students vocalizations under the control of the teachers model and external
reinforcement accomplishes two goals: it puts the teacher in a better position to shape the
students vocal behavior into appropriate speech, and it can teach the student that vocal
sounds can be used to produce reinforcers from the social environment, which may facilitate the development of appropriate social behaviors that replace tantrums and other inappropriate behaviors. Before beginning verbal imitation training, the student should be
able to sit in a chair and attend to an instructor. While generalization between verbal and
nonverbal imitative repertoires should not be unequivocally expected, it may be helpful
for individuals beginning verbal imitation training to have demonstrated mastery of many
nonverbal imitation items, particularly those that involve oral-motor movements.
Verbal imitation training procedures are similar to those used in nonverbal imitation training except that, instead of presenting models of motor actions, they use verbal
models of sounds, words, and eventually sentences. A primary difference in the teaching
procedures for verbal imitation has to do with the way target responses are prompted.
While physical prompting is the primary strategy with nonverbal imitation, only verbal
prompting methods can be used in teaching verbal imitation, since there is no way to
physically prompt a verbal response. A typical verbal imitation teaching trial involves the

sensory matching developmental milestone are at the prelistener level of verbal capability
and do not reliably attend to sensory stimuli. Most children acquire this foundation of
verbal development early in life in the absence of specialized instruction. However, children with certain disabilities may not acquire it without special behavioral developmental
interventions. The sensory matching protocol, outlined in table 3.6, provides children
with rotated multiple-exemplar experiences across five critical sensory modalities. The
rotated exposure to sensory matching experiences provided within this protocol supports
the development of this capability as well as more-complex listener behaviors (Keohane &
Greer, 2005). Greer and Ross (2008) argue that developing a capacity for sameness may
be the fundamental step toward becoming verbal. When a child matches across all senses,
she learns an arbitrarily applicable cross-modal response of sameness.
The objective of this program is to provide children with the capacity for sameness
across different sensory modalities. If children master this protocol, they typically have
the foundation for the abstraction of sameness across sensory stimuli.

1. A child orientating toward adults or other children calling her name

2. A child orientating toward an adult or child in the immediate environment as the first step in a conversational unit
3. A child emitting sustained eye contact with a stimulus relocated by
an adult or other child in the immediate environment
4. A child responding to instructions given by an adult in the immediate environment
5. A child emitting functional self-talk in a play area or other appropriate setting
Materials

TABLE 3.6. The Sensory Matching Protocol:

Matching Across the Senses
Rationale

We use this protocol if adult voices and visual stimuli are conditioned
reinforcers but children are not meeting short-term and long-term
objectives at an adequate rate and do not have the capacity to match
across the senses. Assuming 20-learn-unit instructional sessions, an
adequate rate of learning would be 80 to 120 learn units to mastery
of instructional objectives.

Pre- and Post- The target of this protocol is to accelerate learning rates, just as in the
intervention visual tracking goal. Increased attention is often a collateral effect and
Probes to Test should be measured also.
for Acquisition Pre- and post-probe measures of learn units to criterion, long-term
of the Cusp and short-term objectives achieved, and observing responses associated

with listener and visual sensory modalities (see Pre- and post-probes of
observing responses, below) should be conducted. Probes of noncontextual self-talk (such as palilalia) may also be conducted if this is
a problem for the child.
Pre- and post-probes of observing responses: In a 20-trial format
(a trial should continue for at least 1 second to meet the response
definition criterion) the duration of the following responses should be
measured.
Data collection settings: In one-to-one (e.g., teacher-child), small
group (e.g., 2 to 6 children), and unstructured settings (e.g., play area),
the duration of the following responses should be measured:

Select items and presentation formats that assure that the child can
identify the matching items only through the targeted sensory modality.
Select 2 exemplars across each of the 5 senses. Some suggestions for the
preparation of pairs of sensory stimuli are as follows: (1) auditory sense
(a dog barking and water running), (2) visual sense (picture of a cow
and picture of a house), (3) tactile sense (sandpaper and velvet placed in
separate sacks, one with the exemplar, and the other with the exemplar
and a nonexemplar), (4) smelling or olfactory sense (vanilla or orange
scents in matching containers), and (5) tasting or gustatory sense
(sweetened juice versus water). Each learn unit presentation includes
a correct match exemplar (the positive or target) and 2 incorrect or
negative exemplar stimuli in a matching-to-sample format.

Special Note For the auditory matching component, place a device on the table with
prerecorded sounds of the selected auditory stimuli with accurate or
positive exemplars and one nonexemplar or inaccurate exemplar per
matching-to-sample procedure. Alternate orientation or positions to
ensure that position is not a factor. For the tactile component, a
comparison can be made only when the child is touching the contents
of each sack simultaneously. The child must not be able to see the
tactile stimuli. All instructional programs other than those associated
with self-management and conditioning of preferred interests and
activities should be placed on hold or discontinued until the cusp is
achieved. It is counterproductive to continue unsuccessful instruction.
General
Procedure

Presentation of stimuli across the senses is rotated across 20-learn-unit

sessions. For example, an olfactory learn unit is followed by a tactile
learn unit, and then a gustatory learn unit, followed by an auditory one,
and finally a visual one. Place one exemplar and one nonexemplar on a
table; rotate nonexemplars across instructional presentations. Continue
to use this format across 20 instructional presentations per session.
Rotate the sequence of instruction and the nonexemplars so that the
child does not learn to respond in a particular order, and also rotate
orientation of positive and negative exemplars per learn unit.

TABLE 3.5. Conditioning Print Stimuli on a

Page Protocol: Conditioning Sustained Eye
Contact with Print Stimuli

Pre-test and post-test measures of selected programs should be conducted after each
short-term objective is achieved. When criterion is achieved the childs full schedule of
programs or academic lessons should be resumed and post-test data collected across all
areas of the curriculum.

Conditioning Print Stimuli

Orienting toward instructional materials and the table at which instruction is conducted
is an important early observing response. Additionally, matching tasks provide important educational goals because successful matching performance demonstrates that a child
can visually attend to stimuli, a prerequisite to learning, as well as discriminate among
the relevant properties of environmental stimuli. Basic matching involves responding
to the sameness of identical objects, identical pictures, objects to pictures, and pictures
to objects. More advanced matching may include abstractions such as matching colors,
shapes, irrelevant dimensions (such as a spotted dog versus a black dog), numbers, and
letters. The ultimate goal is for children to acquire generalized matching skills so that
they can match anything without direct instruction, including novel items that they have
never experienced before. Children with significant developmental disabilities often fail to
master both basic and generalized matching skills despite the employment of numerous
remedial techniques such as stimulus or response prompts. We have identified an effective protocol, the conditioning print stimuli on a page protocol, outlined in table 3.5, for
inducing generalized matching in such children. For children at an earlier stage of verbal
development, the visual tracking protocol (Keohane, Greer, & Ackerman, 2006) can be
used to teach looking at three-dimensional stimuli on a table, and as a result, threedimensional visual matching responses may be acquired. However, for other children
we find that print stimuli do not serve as conditioned reinforcers. These children may
acquire basic matching skills with three-dimensional stimuli only and are thus appropriate candidates for the conditioning print stimuli on a page protocol. Our research and
clinical work has shown that when print stimuli and the pictures on pages of books are
established as conditioned reinforcers for observing, children readily acquire generalized
matching skills. In addition, we find that the children are more attentive to in a variety
of settings during and outside instructional sessions.
Observing print stimuli is a critical cusp for a child to have. When a child does not
attend to such stimuli, more complex instructional tasks cannot be targeted. Attention
to stimuli serves as a building block for complex skills, such as joint attention (the capacity to use gestures and eye contact to coordinate attention with another person in order
to share the experience of some object or event in the environment; Mundy, Sigman, &
Kasari, 1994; see also chapters 4 and 13 of this volume).
The conditioning stimuli on a page protocol involves delivering preferred items or
edibles while the child attends to two-dimensional stimuli on a page using the conditioning procedure described earlier in this chapter. The duration of the childs attention to
such stimuli is measured on pre- and post-test probes.

Rationale

Use this protocol if the child fails to attend to print stimuli and has a
high number of learn units to criterion on matching programs.

Pre- and Post- Pre- and post-test probe learn units to criterion on visual (match/
test Trials duplicate) and visual-listener (point/show) learn units to criterion based
on a minimum of 1,000 learn units and 1 criterion.
Pre- and post-test probes of looking at stimuli on a page. These probes
occur following criterion on each short-term objective of conditioning.
The teacher records whether the child looks at a single page of stimuli
for 10 consecutive seconds. Five individual pages are presented 1 page
at a time and these same pages are reserved for post-conditioning probes
only (do not use during conditioning sessions).
Materials

Materials include a variety of 15 to 20 nonpreferred 2-D stimuli

(printed letters, numbers, or pictures) on 8.5-by-11-inch sheets of paper.

Special Note All matching programs and point-to programs or academic lessons are
suspended during the implementation of this protocol. Only return to
these programs when the child meets the long-term objective (LTO) for
this protocol. (The term LTO indicates that the child has acquired the
developmental cusp of conditioned reinforcement for print stimuli.)
General
Procedure

Criterion

Deliver edibles or noninterfering conditioned reinforcers as the child

looks at various pages of stimuli with no observable stereotypy or
passivity. Conduct probe sessions prior to and after each short-term
objective of the conditioning procedure, until the LTO is achieved
during the probes for looking at stimuli on a page.
Conditioning procedure = 90% for 2 consecutive sessions.
Post-probe sessions for looking at stimuli on a page = 80% or 4 out
of 5. This is the LTO for the protocol.

Development of a Capacity for Sameness

Across the Senses
Having the developmental capacity for sameness across the senses helps children advance
to new behavioral developmental cusps. Acquisition of skills that combine the auditory, visual, tactile, olfactory, and gustatory senses is a critical developmental milestone.
Children cannot fully function in a verbal environment without coordination of the
various sensory observations and discrimination skills. Children who have not met the

instructor giving a verbal instruction such as Say ball. For students who engage in
echolalia, or the repetition of vocalizations, the initial instruction of Say may be omitted
or stated with less volume so that the student does not repeat it. Reinforcement is then
provided for the correct imitation of the model. Just as is done with nonverbal imitation, verbal imitation training starts with simple responses such as sounds and progresses
to more complex responses such as words and sentences. Initial verbal imitation training often involves bringing sounds that the student already produces, such as humming
and babbling, under instructional control. Before beginning verbal imitation training, a
teacher may find it helpful to observe the student and record the sounds that he or she
produces repeatedly. Then the teacher can attempt to bring one or more of these sounds
under instructional control by modeling it and providing reinforcement when the student
produces the sound in response to the model.
The following are guidelines for teaching generalized imitation:
1. Choose target behaviors that match the motor or verbal abilities of the student.
In nonverbal imitation training, many individuals find it easier to learn nonverbal imitative behaviors that involve the simple manipulation of objects rather
than those that only involve the movement of body parts. In verbal imitation
training, consider starting with sounds that the student already produces and
bringing them under instructional control.
2. Sit across from the student and have a table available for materials if the target
response requires them.
3. If the target response involves the manipulation of an object, include distracter
objects on the table.
4. Before presenting an instruction, require the student to look at you or at the
materials.

Summary: Teaching Basic Learner-Readiness Skills

Through prompting and differential reinforcement, students with autism and related disabilities can learn to sit, attend to instructional materials, make eye contact with their
instructors, and imitate simple oral and gross motor movements. Often, these skills can
be taught within the context of other instructional programs. These skills, while useful
in their own right, have more importance because they prepare students to learn morecomplex language, social, and intellectual skills, including those that may emerge from a
reinforced history of relational responding, as will be discussed in the coming chapters.

References
Carr, J., Nicholson, T., & Higbee, T. (2000). Evaluation of a brief multiple-stimulus
preference assessment in a naturalistic context. Journal of Applied Behavior Analysis,
33, 353357.
Cooper, J., Heron, T., & Heward, W. (2007). Applied behavior analysis (2nd ed.). Upper
Saddle River, NJ: Pearson Merrill Prentice Hall.
DeLeon, I., & Iwata, B. (1996). Evaluation of a multiple-stimulus presentation format for
assessing reinforcer preferences. Journal of Applied Behavior Analysis 29, 519533.
DeLeon, I., Iwata, B., & Roscoe, E. (1997). Displacement of leisure reinforcers by food
during preference assessments. Journal of Applied Behavior Analysis, 30, 475484.
Doran J., & Holland, J. G. (1979). Control by stimulus features during fading. Journal of
the Experimental Analysis of Behavior, 31, 177187.
Fisher, W., Piazza, C., Bowman, L., Hagopian, L., Owens, J., & Slevin, I. (1992). A
comparison of two approaches for identifying reinforcers for persons with severe and
profound disabilities. Journal of Applied Behavior Analysis, 25, 491498.

5. Give the verbal instruction Do this while simultaneously performing the

motor response for nonverbal imitation. For verbal imitation, give the verbal
instruction Say followed by the sound, word, or sentence
that the student is to imitate.

Graff, R., & Gibson, L. (2003). Using pictures to assess reinforcers in individuals with
developmental disabilities. Behavior Modification, 27, 470483.

6. For nonverbal imitation, provide physical prompting as necessary to help the

student imitate the motor response. In some cases, a second person may be
required to provide prompting during the initial stages of imitation training.

Halle, J. W., Marshall, A. M., & Spradlin, J. E. (1979). Time delay: A technique to
increase language use and facilitate generalization in retarded children. Journal of
Applied Behavior Analysis, 12, 431439.

7. Provide reinforcement for correct imitation of the motor or verbal response

(in general, the more independent the response, the greater the quality of the
reinforcer).

Higbee, T., Carr, J., & Harrison, C. (1999). The effects of pictorial versus tangible
stimuli in stimulus preference assessments. Research in Developmental Disabilities, 20,
6372.

8. Fade prompts as quickly as possible.

Higbee, T., Carr, J., & Harrison, C. (2000). Further evaluation of the multiple-stimulus
preference assessment. Research in Developmental Disabilities, 21, 6173.

9. Test for generalized imitation periodically (after the student has mastered several
targets) by presenting novel targets.

Lovaas, O. I. (1981). Teaching developmentally disabled children: The me book. Austin, TX:
PRO-ED.

For more detailed information about teaching imitation, see chapters 13 and 22 in O.
I. Lovaass book Teaching Individuals with Developmental Delays (2003).

Lovaas, O. I. (2003). Teaching individuals with developmental delays. Austin, TX:

PRO-ED.

Pace, G., Ivancic, M., Edwards, G., Iwata, B., & Page, T. (1985). Assessments of stimulus preference and reinforcer value with profoundly retarded individuals. Journal of
Applied Behavior Analysis, 18, 249255.

CHAPTER 2

Pre- and post-protocol probes of sustained eye contact with stimuli:

A 20-trial probe measuring the duration of sustained eye contact with
a neutral or a nonpreferred stimulus should be completed for each trial
lasting 1 or more seconds.

Peterson, R. F., & Whitehurst, G. J. (1971). A variable influencing the performance of

generalized imitative behaviors. Journal of Applied Behavior Analysis, 4, 19.
Sidman, M., & Stoddard, L. T. (1967). The effectiveness of fading in programming a
simultaneous form discrimination for retarded children. Journal of the Experimental
Analysis of Behavior, 10, 315.
Stoddard, L. T., & Sidman, M. (1967). The effects of errors on childrens performance
on a circle-ellipse discrimination. Journal of the Experimental Analysis of Behavior, 10,
261270.
Taylor, B., & McDonough, K. (1996). Selecting teaching programs. In C. Maurice,
G. Green, & S. Luce (Eds.), Behavioral intervention for young children with autism.
Austin, TX: PRO-ED.
Tessing, J., Napolitano, D., McAdam, D., DiCesare, A., & Axelrod, S. (2006). The effects
of providing access to stimuli following choice making during vocal preference assessments. Journal of Applied Behavior Analysis, 39, 501506.

Criterion for probes: If the child emits 160 (or more) cumulative
seconds of sustained eye contact with the stimulus in a maximum of 20
trials, criterion for the developmental cusp has been achieved.

The Assessment of Basic Learning Abilities

(ABLA) and Its Relation to the Development
of Stimulus Relations in Persons with Autism
and Other Intellectual Disabilities

Pre- and post-probes of observing responses: In a 20-trial format (a

trial should continue for at least 1 second to meet the response definition criterion), the duration of the following responses should be
measured.
Data collection settings: In one-to-one (teacher and child, for
example), small group (2 to 6 children, for example), and unstructured
settings (such as a play area), the duration of the following responses
should be measured:

W. Larry Williams and Marianne L. Jackson,

University of Nevada, Reno

Special Note Programs other than those associated with self-management and an
expanded community of activities and interests (such as looking at
books, playing with toys, manipulating puzzles, and so on) should be
suspended during the implementation of the visual tracking protocol.
We suggest this because the child is not likely to progress with visual
instructional programs at this point, and it is counterproductive to
continue those programs while you are attempting to induce the cusp.
However, programs for conditioning or expanding the childs community of reinforcers or interests can be conducted simultaneously since
these are also reinforcers that increase a variety of observing responses.
General
Procedure

Data Collection: During implementation of the protocol, pairing trials

(of 1 second or more) are recorded as duration in each 20-trial session.
When the child looks away from the stimulus the duration recording is stopped for that trial, and a new trial is begun after the childs
attention is regained. Duration criterion for the long-term objective is
160 or more seconds of sustained eye contact in a 1- to 20-trial session
(in other words, it is possible for the child to reach the 160-second
cumulative objective in a session containing less than 20 trials). Each
20-second trial is an opportunity to test the duration of stimulus
control and a convenient way of delineating sessions.
Pairing procedure: Edibles, tokens, toys, and the like placed under
transparent containers should be used. The child may be prompted to
visually track the preferred item under the container during the first
few pairing trials.
Short-term objectives as tactics: Short-term objectives include incremental increases in targeted duration of sustained eye contact and
interspersal of known items (such as running pairing trials until it
becomes difficult to regain the childs attention). At that point, intersperse known items (for example, have the child respond to the kinds
of instruction he has already mastered, ensuring that you do not lose
the childs attention). In addition, use the preferred item the child
is tracking as a reinforcer for correct responding to the interspersed
trials for previously mastered stimuli). Rate of rotation of stimuli and
number of stimuli rotated (2 or 3) may be used as short-term objectives in this program (for example, stimuli may be rotated 1 to 4 times
at progressively increasing rates of rotation as short-term objectives
become more complex).
Note: The child is not given an opportunity to receive the preferred
item after any pairing trial. We do this because we want only the visual
stimulus control of the object to be the conditioned reinforcer.

Criterion

Two consecutive sessions with 20 (or fewer) trials resulting in 160

cumulative seconds of sustained eye contact with the target stimulus.
All programs should be reintroduced and post-probe data collected
when the child meets criterion.

1. Child making sustained eye contact with a speaker when his name
is called from a distance of 1 to 4 feet
2. Child making sustained eye contact with a speaker when his name
is called from a distance of 5 to 8 feet

Touchette, P. E. (1968). The effects of graduated stimulus change on the acquisition of a

simple discrimination in severely retarded boys. Journal of the Experimental Analysis
of Behavior, 11, 3948.

3. Child looking toward a speaker when the child is given a direction

from a distance of 1 to 4 feet
4. Child looking toward a speaker when the child is given a direction
from a distance of 5 to 8 feet
5. Child looking toward a speaker when another child is spoken to
from a distance of 1 to 4 feet

The objective of this chapter is to acquaint the reader with a practical clinical tool, the
Assessment of Basic Learning Abilities (ABLA). This assessment informs the teacher
or therapist working with children with intellectual disabilities about the appropriate
form and difficulty level of the materials and methods for teaching basic concepts and
communication skills.

6. Child looking toward a speaker when another child is spoken to

from a distance of 5 to 8 feet
7. Child making sustained visual contact as an adult rearranges the
childs materials on the desk
8. Child making sustained visual contact as an adult removes the
childs materials from the desk

Discrimination Learning
Learning to discriminate between relevant stimuli in the environment is crucial to learning many functional skills, including communication and social skills, and is an assumed
skill in many psychometric tests. Matching to sample (MTS) is one of the most commonly used paradigms for teaching and assessing discrimination skills and is also the
training protocol commonly used to assess relational responding for communication skills,
elementary reading, and equivalence relations (see chapter 8 of this volume). Therefore,
if individuals do not demonstrate the ability to match to sample, it renders them basically untestable on many psychometric tools. MTS is also frequently used in many classroom and preacademic training settings and has been instrumental in communication

9. Child looking toward an adult entering the room who is speaking

10. Child making sustained visual contact with an adult moving about
the room who is not speaking
Materials

Use 2 or 3 identical transparent containers (cups or other containers).

Vary size and shape of identical sets for each presentation. A variety of
neutral items for the pre- and post-probes (paper clips, unfamiliar shapes,
and other items) and preferred items (edibles, tokens, toys, and the like)
that can be placed under the target container should be available. A stopwatch is used to record duration of sustained eye contact.

above under Pre- and Post-test Probe Trials. In the pairing segment,
edibles are typically paired with listening to adult voices until the child
listens with no observable stereotypy (stereotypy is a competing reinforcer). During the pairing intervals, 2 and 3 pairings of edibles should
be rotated (the number stays the same as the pairing intervals graduate
from 5 seconds to 10 seconds, then 15 seconds, and so on). No reinforcement procedures are used during test trials.
Sessions are typically 5 minutes in duration and whole interval
continuous 5-second intervals constitute measurement of the students
progress in achieving criterion on the conditioning intervention.
Criterion

90% of 5-second whole interval recordings (ninety 5-second intervals)

over two 5-minute consecutive sessions. Criterion for meeting the
test of conditioned reinforcement is 90% + intervals for 2 consecutive
5-minute sessions with no observation of stereotypy or passivity. The
voice is a conditioned reinforcer when the child will touch the disk or
hold down the button continuously for 90% of the observation intervals recorded in 5-second intervals within a 5-minute session.

This protocol has been repeatedly shown to be effective in increasing a childs listening to adult voices (Greer, Keohane, & Delgado, 2006), conditioning listening to specific
music (Greer, Dorow, & Hanser, 1973; Greer, Dorow, Wachhaus, & White, 1973) and
acquisition of more complex verbal, academic, and social skill sets (Tsai & Greer, 2006).

visual stimulus is selected out is a measure of the reinforcement by that stimulus of the
childs operant observing behavior. We frequently find that when a child is having difficulty acquiring prelistener verbal capabilities, the problem is related to a lack of attending behaviors associated with specific and general observation of the environment. The
objective of the visual tracking protocol, outlined in table 3.4, is to condition sustained
observation of visual stimuli by pairing unconditioned or conditioned reinforcers (such
as edibles and toys) with the childs attention to visual stimuli (Keohane & Greer, 2005).
In this way, we transfer the reinforcement control from the unconditioned or conditioned
reinforcement to the previously neutral stimulus (in other words, attending to the visual
stimuli that is now a reinforcer for the operant observing response). See the introduction
to Tsai and Greer (2006) for references to the extensive laboratory research on conjugate
reinforcement.
Prelisteners who attend to visual stimuli inconsistently, do not imitate teacher modeling, do not match to sample, do not follow basic directions, and fail to meet short-term
and long-term curricular objectives are candidates for the visual tracking protocol. Preand post-test measures of responding to learn units across all areas of the curriculum,
and the observing responses associated with sustained eye contact to stimuli and other
individuals in the environment, are measured before and after mastery of the protocol.

TABLE 3.4. Visual Tracking Protocol: Observing 3-D

Tabletop Stimuli as Conditioned Reinforcement
Rationale

Conditioning Observing Visual Stimuli

An early challenge for those who work with children with severe language delays is
ensuring that the children visually attend to relevant environmental stimuli. Many of
the children we work with do not make sustained eye contact with other individuals or
stimuli in the environment. Sustained attention to visual stimuli is a critical early step
in language acquisition, as these children may otherwise be unable to develop correspondence between what they hear and what they see. Prior research reported that visual
training may actually enhance tactile discrimination skills in young children (Krekling,
Tellevik, & Nordvik, 1989), and that reinforcement in the form of music or other preferred activities may improve visual tracking skills in infants (Darcheville, Madelain,
Buquet, Charlier, & Miossec, 1999). Because of our many years of work with children
with disabilities, we knew that language acquisition was hindered when the children did
not make sustained eye contact with stimuli in the environment. As a result we designed
a protocol using a form of conjugate reinforcement to address the problem. Collier and
Bitetti-Capatides (1979) defined conjugate reinforcement as continuous reinforcement
during which the intensity of the reinforcement is directly related to the intensity of the
response (for example, in vigor, rate, or duration). For example, the longer a child gazes
at a colorful toy, the longer the duration of the reinforcer for gazing at the toythe
sight of the colorful toy itself). That is, the degree to which the childs attention to the

This protocol should be implemented if a child does not attend to

visual stimuli or look at adults or children in the environment and
emits low numbers of correct responses across visual matching programs and other curricula requiring attention to visual stimuli. See
Greer and Ross (2008) for prerequisites.

Experimental These probes are measures of the childs rate of learning instructional
Probes
objectives that entail visual observing. If 3-D visual stimuli are conditioned reinforcers for observing responses, the child will learn at a significantly faster rate than he would if the stimuli do not reinforce looking.
Pre- and post-protocol probes of total learn units to criterion, visual
(e.g., match/duplicate) learn units to criterion, and visual-listener (e.g.,
point/show) learn units to criterion based on a minimum of 1,000 learn
units and 1 criterion per category should be completed.
Pre- and post-protocol probes of eye contact and visual tracking of
items and individuals in three selected environments (one-to-one, small
group, and unstructured) should be completed using duration recording
of each trial lasting 1 or more seconds. This could be done by starting
timing when the child begins observing the stimulus, stopping timing
when the child stops sustained observing of the stimulus, regaining the
childs attention, and beginning another duration recording trial.

and language training. Teaching an individual to match to sample appears to be basic for
increasing the array of basic concepts that an individual might learn.
There are a number of types of discriminations that are relevant and even vital to
many educational skills. These discriminations could be described as existing on a continuum of difficulty or complexity. Nonrelational discriminations may represent some of
the least complex discriminations, and arbitrary conditional discriminations could be said
to represent a more complex type of discrimination.
Nonrelational discriminations involve a simple, simultaneous discrimination that does
not require the presence of a sample stimulus. A basic relational discrimination is said to
have occurred when a learner comes to respond to the presence of a given environmental
event or stimulus (discriminative stimulus) and does not respond in the absence of that
event or stimulus. For example, a simple, nonrelational discrimination would be learning that in the presence of a plate the correct response is to put it in the dishwasher. In
the case of a simple discrimination, this response will occur regardless of whether or not
the plate is dirty and needs to be washed. The presence of the plate itself functions as the
stimulus for the response of placing it in the dishwasher.
In relational responding, the individual must attend to two stimuli and respond on
the basis of their relationship (in MTS, the sample and correct comparison). A conditional
discrimination requires that the function of the comparison stimuli change from trial to
trial depending upon the sample stimulus, such that a given stimulus is presented over
trials as both the correct and incorrect comparison. This type of discrimination represents
a logical if-then rule. Following on the previous example of a simple discrimination, this
would represent a conditional discrimination if the individual responded by only putting
a used or dirty plate (sample stimulus) in the dishwasher but not putting a clean plate
in the dishwasher. A simpler example of a conditional discrimination would be matching a red sample to a red comparison and matching a blue sample to a blue comparison
stimulus.
Conditional discriminations may occur within and across any sensory modality,
including visual, auditory, olfactory, and tactual. Moreover, they may involve the features of formal physical similarity, referred to as nonarbitrary conditional discriminations, or they may be formally dissimilar and thus be completely arbitrary conditional
discriminations.

largely because the tests are short and hence provide few opportunities for learning to
occur throughout the assessment. In addition, research shows that levels at which the
learner initially fails are very difficult for them to ultimately master, often requiring a
large number of trials (Meyerson, 1977).

ABLA Procedures and Materials

The ABLA consists of six levels of testing, each examining different discrimination
abilities. Testing at each level begins with a demonstration trial, a guided trial, and a
practice trial, and reinforcement is provided for every independent correct response. The
passing criterion for each level is eight consecutive independent correct responses; the
failing criterion is eight cumulative errors on a given level. Each progressive level presents
a more complex type of discrimination and builds on the skills tested at previous levels.
Testing is terminated upon the failure at any level, because research has shown that an
individual who fails a particular level will not pass higher levels (Kerr, et al., 1977).
The materials used in the ABLA are a large yellow can of approximately 6 inches in
diameter and 7 inches in height, a large red box approximately 5.5 by 5.5 by 4 inches, a
small yellow cylinder approximately 1.5 inches in diameter and 3 inches in height, a small
red cube with approximate dimensions of 2 by 2 by 2 inches, and two small, shapeless
pieces of gray foam approximately 2 inches in diameter.
All correct responses should be reinforced with items that are appropriate for that
individual. These should be assessed prior to testing. Incorrect responses should result
in a neutral no from the tester and no other differential consequence, followed by the
prompting hierarchy outlined in Figure 2.1, which depicts the operationalization of each
instructional trial. It provides descriptions of the appropriate consequences for correct
and incorrect responses and the prompting hierarchy as necessary for error correction.
Table 2.1 delineates the relevant instructional materials, instructions given to the learner,
and response requirements for each level. By using the materials and procedures listed in
table 2.1 and the flowchart in figure 2.1, a variety of individuals should be able to use the
ABLA as a simple and relatively quick assessment tool.

Level 1

The ABLA and Conditional Discriminations

The Assessment of Basic Learning Abilities (ABLA), which was developed by Kerr,
Meyerson, and Flora in 1977, tests an individuals ability to form some of these various
types of discriminations. The ABLA is typically administered to individuals with some
degree of intellectual disability and examines the ease or difficulty with which an individual can learn one simple motor task and five two-choice discrimination tasks. The
ABLA is a short test that can be administered relatively easily and quickly using a small
number of stimuli.
Performance on the ABLA provides information that allows inferences to be drawn
regarding individuals prior histories with such tasks. These inferences can be drawn

In level 1 the tester presents one of the two containers, either the yellow can or the
red box, on the table in front of the individual. Both the tester and the individual being
tested have a piece of gray foam and the tester gives the instruction Do this while
placing his or her piece of foam in the container. The individual responds correctly when
he or she places his or her piece of foam in the same container. This procedure is then
conducted with the other container so as not to develop differential histories with respect
to each. Passing and failing criteria are as described above.

Level 2
In level 2 the tester presents both the red box and the yellow can on the table, with the
yellow can on the individuals left side and the red box on the right. The individual being

tested is given a small piece of gray foam and is provided with the standard demonstration
trial, a guided trial, and a practice trial before being asked for an independent response.
The instruction given is Where does it go? and a correct response is the placement of
the foam in the yellow container, on the individuals left side. This level tests the individuals ability to learn a positional discrimination.

TABLE 2.1. ABLA Materials and Methods

Level

Level 3
Level 3 is similar to level 2, with the exception that the positions of the two containers are randomly alternated between trials. During any given trial the correct response is
the placement of the piece of gray foam in the yellow can, regardless of its position on the
table. This tests the individuals ability to learn a simple discrimination.

Level 4

Level 4 presents the first test of conditional discriminations. Both the yellow can
and the red box are present on the table in front of the participant and their position
alternates randomly between trials. The participant is given either the small yellow cylinder or the small red cube and the instruction Where does it go? After being provided with a demonstration trial, a guided trial, and a practice trial on both stimuli, the
individual responds correctly by placing the small red cube into the large red box, or
the small yellow cylinder into the large yellow can. The presentation of the small yellow
cylinder or small red cube alternates randomly between trials. This level tests the individuals ability to learn a visual-visual conditional discrimination (sometimes referred to
as a quasi-identity match, since the sample and correct comparison are formally similar
but not identical).

Level 5
Testing in level 5 introduces cross-modal discriminations. These discriminations
involve samples and comparisons from different modalities, specifically visual and auditory, whereas all previous levels tested discriminations within the visual modality. In level
5 both the yellow can and the red box are placed on the table and their positions do not
vary during this level. The individual is given a small piece of gray foam and asked to put
it in the red box (said in a rapid or staccato fashion) or put it in the yehlllloooow
caaan (said in a slow, drawn-out fashion). This assesses the individuals ability to learn a
simple auditory-visual discrimination.

Level 6
Level 6 is similar to level 5 with the exception that the yellow can and red box alternate positions randomly on each trial. The instructions given are the same. This assesses
the individuals ability to perform a conditional auditory-visual discrimination. As stated
previously, testing is terminated upon the failure of any particular level.

Materials

Instruction

Correct Response

 Yellow can or red box in  Do this (as the  Individual picks

center of table (but only
assessor picks up a up his or her
one of them on any given
piece of gray foam piece of foam
trial).
and places it in
and places it in
the container on the container on
 Two pieces of gray foam
the table, either
the table.
(one in front of the assessor
the yellow can or
and one in front of the
the red box)
individual being tested).
 Individual
 Yellow can and red box,  Where does
takes the piece
with yellow can positioned it go? (as the
of foam and
to the left of the individual piece of foam is
presented to the
places it in the
being tested and the red
yellow can,
box to the right. Both will individual)
which is always
be at an equal distance
positioned to his
from the individual.
or her left.
 Positions of the yellow can
and red box do not change
over trials.
 One piece of foam is
handed to the individual to
be tested.

 Yellow can and red box,  Where does

 Individual takes
both positioned in front of it go? (as the
the piece of
the individual and at equal piece of foam is
foam and places
distances.
presented to the
it in the yellow
individual)
can (regardless
 At this level the left and
of its left or
right positions of these
right position on
stimuli will alternate
the table).
randomly on each trial,
so starting position is not
important.
 One piece of foam is
handed to the individual
being tested.

child does not meet the 90 percent criterion for the five-minute free operant test of listening to the recorded voices, we increase the duration of the pair and test intervals (such as
ten seconds, fifteen seconds, and twenty seconds), until the child meets criterion.

2. Child orienting toward a speaker when her name is called from a

distance of 5 to 8 feet

TABLE 3.3. Conditioning Listening to Adult

Voices Protocol: Conditioning Adult Voices as
Reinforcement for Observing Responses

4. Child orienting toward a speaker when the child is given a direction

from a distance of 5 to 8 feet

Rationale

Pre- and
Postintervention
Probes to
Test for the
Acquisition
of the Cusp

3. Child orienting toward a speaker when the child is given a direction

from a distance of 1 to 4 feet

5. Child orienting toward a speaker when another child is spoken to

from a distance of 1 to 4 feet
6. Child orienting toward a speaker when another child is spoken to
from a distance of 5 to 8 feet

This protocol is indicated if a child does not orient toward adult voices
and/or look at speakers, particularly those holding sources of reinforcement. Be certain the child does not have a major hearing deficit before
attempting to teach this stimulus control. If voices do not select out
or attract the childs attention, the child is unlikely to be prepared to
discriminate vowel-consonant sounds and other aspects of speech that
come to have listener and speaker effects. If adult voices are conditioned reinforcers for observing responses, the child will learn at a
significantly faster rate.
Twenty experimental probe trials (no consequences) should be completed using duration recording of each trial lasting one or more
seconds. These probe trials should consist of a variety of novel opportunities for the child to respond to an adults presence (for example,
the child turns toward an adult when her name is called, looks toward
an adult entering the room, looks toward an adult speaking to a child
nearby, or looks toward an adult rearranging the childs environment,
such as moving a toy or other tabletop materials; please see Pre- and
Post-probes of Observing Responses below), measured in three selected
environments (one-to-one, small group, and unstructured settings).
Pre- and post-protocol probes of total learn units to criterion across
subject area lessons (such as match/duplicate, point/show) based on a
minimum of 1,000 learn units and 1 criterion per category should be
completed.
Pre- and post-probes of observing responses: In a 20-trial format (a
trial should continue for at least 1 second to meet the response definition criterion), the duration of the following responses should be
measured:
Data collection settings: In one-to-one (such as teacher and child),
small group (for example, 2 to 6 children), and unstructured settings
(such as a play area), the duration of the following responses should be
measured:
1. Child orienting toward a speaker when her name is called from a
distance of 1 to 4 feet

7. Child orienting toward an adult rearranging the childs materials on

the desk
8. Child orienting toward an adult removing the childs materials from
the desk
9. Child orienting toward an adult who is entering the room and
speaking.
10. Child orienting toward an adult entering the room who is not
speaking
Materials

Use nursery rhymes or selections from childrens books recorded by the

childs mother, teacher, and someone the child does not know. Tapes or
CDs should be 5 minutes in duration and should not include singing.
Have 6 or more recordings available that are rotated in equal measure
across speakers.

Special Note Academic lessons other than those associated with an expanded community of activities and interests (such as looking at books, playing
with toys, manipulating puzzles, and so on) are suspended during the
implementation of the protocol. A multiple probe format is used.
Conditioning Use a tape recorder or other recording device that reproduces the
Procedure selected voices when the button is manually held down by the
child and automatically stops when the child takes her hand off the
button. Electric switches from speech therapy catalogues can be used.
Alternatively, you may substitute a laminated circle or square that will
act as a simulated on/off button or switch for the child to touch while
the teacher controls the progress of the recording device. (If the child
has her hand on the simulated disk or switch, the tape is played; if the
child takes her hand off the disk or switch, the tape is stopped immediately.) Note that touching the disk or switch, is the most direct measure
of the auditory observational response, and this allows us to determine
if the child is listening or not listening. If the child emits stereotypy,
the tape should be stopped as well. A pair-test conditioning procedure
with pre- and post-tests of the childs orienting to voices is described

TABLE 3.1. Prelistener: Very Early

Verbal Developmental Cusps
Verbal Developmental Cusp

Protocol

Adult voices as conditioned reinforcers

Conditioning listening to adult voices

Conditioned reinforcement for visual

stimuli

Conditioning visual stimuli as

reinforcement for tracking/print

The capacity for sameness across the

senses

Cross-modal sensory matching

TABLE 3.2. Listener-Speaker:

Early Developmental Corequisite
Verbal Developmental Corequisite

Protocol

Imitation

The mirror procedure

Prelisteners are entirely dependent upon others for everything in their lives. Entrance
into the social community is not possible. Early listeners and early speakers are able to
participate in some aspects of the social community. When children reach speaker-asown-listener levels of verbal capability, they are able to participate in and contribute to the
social community in more comprehensive and independent ways.

The Development of Very Early Observing Responses

Responding to human voices appears to be an important component of childrens acquisition of both listener and speaker repertoires (Pelez-Nogueras, Gewirtz, & Markham,
1996) and, not surprisingly, occurs very early in typically developing infants. Novak and
Pelez (2004) found that even newborn infants are able to discriminate human language
sounds. Since language is arbitrarily defined (for example, consonant-vowel combinations are arbitrarily given meaning), the individuals responding must come under control
of the acoustical properties of speech. DeCasper and Fifer (1980) conducted a study in
which newborns could occasion the playing of a recording of either their mothers voices
or the voice of an unknown female by sucking on a nonnutritive nipple at differential
rates. Once the response requirement was established, it was found that the infants caused
the recordings of their mothers voices to be played more frequently than the recording
of the unknown females voice, suggesting that certain voices may become conditioned

reinforcers early in life. We propose that the mothers voices selected out the attention of
the children because they were conditioned reinforcers for observing responses.
Most typically developing infants begin to differentiate between consonant-vowel combinations, such as pa and ba, within their first month (Eimas, Siqueland, Jusczyk, &
Vigorito, 1971); in fact, vocal imitation of phonetic units appears to play a critical role
in infants acquisition of language (Kuhl & Meltzoff, 1996). For children with significant
language delays, however, adult voices do not serve as reinforcers for listening and, as a
result, prelistener skills may not develop. If voices do not select out attention, discriminations of the pa and ba sounds are not likely to occur. Field (1987) found that children at
higher risk for developmental disabilities did not develop positive social interactions through
typical mother-infant play, and they needed more intensive stimulation than did typically
developing children before responding positively to a caregivers touch and voice. Our work
suggests that the attainment of conditioned reinforcement for listening to adult voices is one
of the earliest capabilities associated with language development and social functioning.
Children who are selected for the conditioning listening to adult voices protocol,
outlined in table 3.3, are assessed to be prelisteners. They do not typically respond to
visual or auditory stimuli in the environment and as a result cannot progress toward their
curricular goals. Pre- and post-test measures of the childrens levels of responding to learn
units across all areas of the curriculum are taken before and after mastery of the protocol,
as are measures of nonfunctional self-talk. Observing responses associated with sustained
eye contact and observation of the activities of other individuals in the immediate environment are also measured before and after mastery of the protocol. The measures of the
numbers of learn units required in order to master components of the curricula are tests
of the rate of learning.
In this chapter, for the protocols we describe based on a pair and test trial format (conditioning listening to adult voices and conditioning print stimuli), we use the Pavlovian
second-order conditioning procedure. The teacher delivers edibles and noninterfering
unconditioned reinforcers (such as edibles) or conditioned reinforcers (such as praise or
a token if they are truly conditioned reinforcers) when the child is attending appropriately to the target item or activity. Data are recorded using a pair and test trial format.
Initially, the first short-term objective begins with a five-second pair trial interval, and
when completed successfully it is followed by a five-second test trial interval. During the
pairing trial, the teacher delivers a reinforcer either two or three times contingent upon
the childs emitting the target behavior. The trials of two and three pairings are alternated. However, if the child does not engage in the target activity or item and/or emits
stereotypy (repetitive or ritualistic movements) or passivity at any time during the pairing
trial, the pairing trial is immediately restarted. The test trial begins only when the pairing
trial is completed without stereotypy or passivity.
During the test trial, the teacher records whether or not the child engages in the
activity or item according to the definition of the target behavior for the duration of the
interval used (five-, ten-, or fifteen-second intervals or longer). No reinforcement is delivered during the test trial. Whole interval recording is used, in which a correct response is
recorded if the child engages appropriately for the entire interval. If at any point the child
is not engaging in the activity or item appropriately, an incorrect response is recorded,
the test session is ended immediately, and the next pairing interval is started. Data are
recorded and graphed out of twenty for the number of correct test trials. Generally,
criterion is set at 90 percent for two consecutive pair and test trial sessions. When the

 Yellow can and red box,

both placed in front of
the individual at equal
distances.
 Again, at this level their
positions will alternate
randomly.

 Yellow can and red box

both placed in front of
the individual at equal
distances.
 Both stay in the same
position on each trial.
 One piece of gray foam
is placed in front of the
participant.

 Yellow can and red box

both placed in front of
the individual at equal
distances.
 The positions of the
containers alternate
randomly on each trial.
 One piece of gray foam
is placed in front of the
participant.

When I say (insert appropriate instruction

for level), you do this (demonstrate correct
response).

Guided trial:

Now lets do it together.

Give instruction for that level, then take the
individuals hand and guide him or her
through the correct response.

 When presented
with the red
cube the correct
response is to
place it in the
red box.

 One small yellow cylinder

and one small red cube
(only one presented to
the individual on any
given trial).

Demonstration trial:

 Where does it  When presented

go? (as either the with the yellow
yellow cylinder
cylinder, the
or red cube is
correct response
presented to the
is to place it in
individual)
the yellow can.

Independent trial:

Now you try. Give instruction for that level

and then allow 3-5 seconds for response.

 Put it in the  Individual places

yehlllloooow
the foam in the
caaan, said in a
container that
slow voice with
corresponds to
a hand covering
the instruction.
the mouth.

No response or incorrect

Correct response

Begin testing

 Put it in the
red box, said
quickly with a
hand covering
the mouth
 Put it in the  Individual places
yehlllloooow
the foam in the
caaan, said in a
container that
slow voice with
corresponds to
a hand covering
the instruction.
the mouth.
 Put it in the
red box, said
quickly with a
hand covering
the mouth

Give instruction for level and

allow 3-5 seconds for response

No response or incorrect

Score as incorrect

Correct response

Score as correct

Demonstration trial
Level is passed after 8
consecutive correct
responses

Guided trial
Independent trial

Begin next level

No response or incorrect

Level is failed after 8 cumulative errors

Figure 2.1. ABLA flowchart.

Testing ends

Common Problems and Suggested Strategies

Given the range of individuals who may be assessed using the ABLA, there are some
common problems that may arise. Most of these will occur during the earlier stages of
testing and are dealt with before higher levels of testing. Furthermore, it is often the case
that individuals who progress to testing on higher levels (such as levels 4, 5, and 6) have
experienced a variety of similar training or testing situations, so these issues may not
arise.
In the early stages of testing, there are three common problems. The first is that individuals may not respond at all to the materials or the instructions given. In this situation,
the assessor should implement the prompting hierarchy described in the flowchart. The
assessor will first provide a demonstration of the correct response, guide the individual to
make the correct response, and then repeat the instruction and provide an opportunity
for the individual to perform the response independently. If the individual responds correctly, the response will be reinforced appropriately and the next trial will begin. If the
individual does not respond correctly or fails to respond, the assessor will record this as
one incorrect response and repeat the prompting hierarchy. Each time the final step in
the prompting hierarchy is implemented (in other words, an opportunity to make an
independent response), a correct or incorrect response will be recorded. This cycle will
continue until the passing or failing criteria are met.
Another common problem encountered in the earlier stages of assessment occurs
when individuals throw materials or try to hand them to you. Throwing of materials
should be considered an incorrect response, and be followed by a neutral no, rapid
replacement of materials, and implementation of the described prompting hierarchy. The
assessor should not directly address the throwing of materials in any other way. In the
event that the individual tries to give the assessor the materials, the assessor should not
accept the materials but should, again, respond with a neutral no and implementation
of the prompting hierarchy.
One final problem often encountered is that the individual demonstrates a position
bias, meaning that he or she constantly responds to any stimulus presented in a specific
position (the right or left side) regardless of the task or instructions given. In addition,
the individual may respond to one of the stimulus items and not the other. Although
this may simply reflect a general lack of ability to form discriminations beyond level 2
or 3 (positional discrimination, or simple discrimination), the effect may be reduced or
avoided by using a simple procedural intervention early in testing. During level 1 it may
be advantageous to use one container (such as the yellow can) for some trials and the
other stimulus (such as the red box) for other trials. Changing the container used may
help prevent overselectivity of one container over the other in later levels of testing.
Although these are only a few examples of the possible errors that individuals may
display throughout testing, they represent some of the more common areas of difficulty experienced. The strategies used to address these problems can also be applied to
almost any error that occurs. In general, re-presenting the trial and prompting a correct
response will be required. Assessors should be sure to record all incorrect responses and
independently made correct responses throughout these procedures.

A Predictive Hierarchy of Discrimination Abilities:

ABLA Findings
When Kerr, Meyerson, and Flora developed this test in 1977, they administered it to
117 individuals. Rising out of this project were a large number of research projects that
began to investigate the nature of these learning abilities and the practical implications of
them (for reviews see Martin & Yu, 2000, and Martin, Yu, & Vause, 2004). One of the
earliest findings of ABLA research was that the test discriminations were hierarchically
ordered according to their difficulty; the findings showed that failure at a particular level
predicted failure at higher levels of the hierarchy. Another finding of the 1977 study by
Kerr, Meyerson, and Flora was that failure at a level occurred relatively quickly and that
failed levels were extremely difficult to teach. Furthermore, this hierarchy of discriminations and the difficulty with which they can be learned has also been found to hold for
typically developing children.

ABLA as a Predictive Assessment

Research has shown that the ABLA not only predicts an individuals ability to make
particular types of discriminations but is also highly predictive of success on educational,
prevocational, and vocational tasks (Stubbings & Martin, 1995). In fact, performance
on the ABLA was found to be a more accurate predictor of success on a variety of training tasks than the assessments of experienced teachers with direct knowledge of a given
individual.
Many studies have investigated the relationship between the ABLA levels and language
abilities. More specifically, it has been shown that individuals who passed ABLA levels
2 and below but failed higher levels typically had no formal verbal abilities. Individuals
who passed up to level 4 but failed higher levels were typically able to communicate
using single words or sounds. Language skills involving two or more words typically only
occurred in individuals who passed level 5 or 6 (Ward & Yu, 2000).
More recently, Richards, Williams, and Follette (2002) examined performance on
the ABLA and the communication, daily living, and social skills subdomains of the
Vineland Adaptive Behavior Scale (Sparrow, Balla, & Ceccetti, 1984). The Vineland
Adaptive Behavior Scale is a broad measure of an individuals adaptive functioning in
these areas. They reported a positive correlation between these tests but found that the
Vineland was not sensitive to differences between individuals below ABLA level 6. Such
individuals showed low age-equivalent scores on the Vineland. Moreover, they found that
the Wechsler Adult Intelligence Scale-Revised (WAIS-R) not only failed to differentiate
individuals below ABLA level 6 but also scored them as untestable. This research suggests
that the ABLA may be a more sensitive and predictive assessment for individuals who
have severe impairments.
Marion and colleagues (2003) found that ABLA performance was a better predictor
of performance on assessments of verbal operants (mands, tacts, and echoics; Skinner,
1957) than diagnosis. Furthermore, their results suggested that ABLA level 6 may be a
prerequisite or bridging task for teaching such verbal operants.

skills, are able to participate in the social community to some extent, can follow a number
of directions, and are less dependent on others for their everyday needs.
We propose that the emergence of imitation through observation, conditioned reinforcement for listening to voices, looking at stimuli and print, and matching stimuli across
the senses or across sensory modalities may be prerequisites for the development of observing responses as related to early language acquisition. Such early language acquisition
occurs across repertoires of listener (for example, a child who follows simple directions),
speaker-listener (for example, a child with vowel-consonant auditory discrimination skills
that result in the production of speech), speaker-as-own-listener (for example, a child who
is able to speak and listen to himself or herself, as in thinking), and cross-modal capacity for sameness (for example, a child who discriminates what is the same and what is
different across sensory modalities).
We have been on an applied behavior analytic journey of sorts, beginning with
Skinners theoretical framework of verbal behavior (1957) and the recent expansions of
that theory (Greer & Keohane, 2005; Hayes et al., 2001; Horne & Lowe, 1996). Along
the way we have incorporated theories, research, and practices related to the basic science
and, when relevant, the infant developmental literature and animal social learning theory.
The cumulative body of literature was very useful to us in our development of instructional protocols for children with significant language delays due to autism or other
developmental disorders. These protocols are based on a progression of complex language
functions or cusps (Baer, 1983; Hart & Risley, 1999; Premack, 2004; Rosales-Ruiz &
Baer, 1997). We also reviewed evidence that nonhumans could be taught certain noncomplex features of language (Epstein, Lanza, & Skinner, 1980, 1981; Premack & Premack,
2003; Savage-Rumbaugh, Rumbaugh, & Boysen, 1978), and that teaching, as distinct
from acquiring repertoires based on modeling, is unique to the human species (Premack,
2004). As we moved forward, it became increasingly clear that listener, speaker-listener,
and speaker-as-own-listener repertoires make complex verbal behavior possible and are
unique to humans.
As part of the process of developing a comprehensive systems-based behavior analytic
approach to teaching and learning over the last twenty-five years, the Comprehensive
Application of Behavior Analysis to Schooling (CABAS) team and colleagues in the
applied behavior analysis programs at Columbia University Teachers College have compiled instructional components based upon new research findings. The CABAS system
includes empirically based curricula and protocols for teaching new operants, the training
and motivation of teachers and other school professionals, and the design of functional
curricula (Greer, 2002). Our work has also allowed us to identify and remediate missing
developmental cusps in children with autism and related developmental disabilities.
Inherent in this system is the conceptualization of learn units, or instructional presentations that provide yoked or interlocking contingencies between teacher and child. Learn
units include (1) the need to know (for example, a motivating condition; (2) the childs
attention to the relevant antecedent stimulus, such as the teachers instructions; (3) an
opportunity for the child to respond; (4) reinforcing consequences immediately following correct responses; and (5) noncoercive corrections that immediately follow incorrect
responses and require the child to repeat the correct response (see Greer, 2002, and Greer
& Ross, 2008, for the extensive research base).
Our early-language training programs focus on self-awareness, or the ability to observe
ourselves, an essential component of language acquisition. When children begin to respond

to adult voices, make sustained eye contact with visual stimuli, learn to imitate through
observation, and match stimuli across the senses, they demonstrate early evidence of selfawareness. They are able to distinguish between themselves and others in their environment. This distinction is elemental to all observing and producing responses and is the
basis for participating in such culturally evolved activities as music, art, or language.

Building Listener Literacy for Children with

Language Delays
In our initial efforts to bridge the educational gap for children with significant communication delays we focused on speaker repertoires, following Skinners lead (1957), or what
is often traditionally described as expressive language. It was a logical approach and was
closely tied to Skinners emphasis on the speaker, since the children we worked with were
frequently unable to speak. However, when the children began to display some speaker
behavior, it was often context specific, scripted, and directly taught. Even though our
assessments focused on both speaker and listener repertoires, including listener skills that
involved responding appropriately to the language of others (as in orienting toward or
identifying environmental stimuli labeled by others and complying with spoken instructions [Skinner, 1957]), we were much better able to measure and produce responses associated with speaking skills.
As a result of our research and our success (as well as our lack of success) with individual children, we reconsidered Skinners comments on the role of the listener. Skinner
described the relationship between the speaker and the listener as follows: When the
listener looks to the speaker for an extension of his own sensory capacities, or for contact
with distant events, or for an accurate characterization of a puzzling situation, the speakers behavior is most useful to him if the environmental control has not been disturbed
by other variables (Skinner, 1957, p. 418). This description provided us with a way to
reconceptualize the components of early-language development and design protocols for
assessment and intervention related to listener skills among children with autism and
related disabilities.

Classifying Childrens Prerequisite Foundations of

Verbal Behavior as Listener Developmental Cusps
As soon as we identified some of the capabilities that children needed in order to progress, these were organized into a hierarchy of verbal developmental milestones or cusps
(Greer & Keohane, 2005). This facilitated the identification of prerequisite skills and
the design of protocols to provide a means for children to advance across these cusps.
Compliance with adult instructions and the establishment of the teacher as a source of
conditioned reinforcement are prerequisites for the three earliest cusps and corequisite
cusp listed below.
Shown in tables 3.1 and 3.2 are listings of prerequisite and corequisite developmental
cusps and the instructional protocols used to establish them.

CHAPTER 3
Observing Responses: Foundations of
Higher-Order Verbal Operants
Dolleen-Day Keohane and Jo Ann Pereira Delgado,
Columbia University Teachers College and CABAS;
R. Douglas Greer,
Columbia University Graduate School of Arts
and Science Teachers College

Observing responses associated with listener and speaker repertoires is the foundation of
certain aspects of early language. Observing responses consist of the operant responses of
looking, listening, tasting, smelling, and touching. The observing operants are selected
out by the consequences that reinforce observation, and the stimuli that reinforce them
are established by reinforcement conditioning processes. These observing responses and
their reinforcers lead to the development of more complex behaviors (Donahoe & Palmer,
2004; Greer & Ross, 2008). Observing responses are critical to production responses,
both of which are inherent in a variety of cultural practices, including art, music, and language. We have identified some of the subcomponents of language as a result of working
inductively toward a hierarchy of verbal development (Greer & Keohane, 2005; Greer &
Ross, 2008), drawing on Skinners verbal behavior theory and extensions of that theory
(Hayes, Barnes-Holmes, & Roche, 2001; Horne & Lowe, 1996; Skinner, 1957, 1989). In
this chapter we will concentrate on prelistener and early-listener capabilities, because they
lead to the emergence of new verbal capabilities, or cusps, in children with autism and
other developmental disabilities. Prelisteners are defined as children who do not observe
their environment, do not participate in the social community, and are completely dependent on others for their very survival. Early listeners are children who have basic o bserving

Further Refinements of the ABLA

The discriminations tested at levels 5 and 6 are arbitrary conditional discriminations,
but they are cross-modal because they involve auditory and visual stimuli. Until Sakko,
Martin, Vause, Martin, and Yu (2004) developed such a testing level, there was no existing test of arbitrary conditional visual-visual (or within modal) discriminations.
The visual-visual nonidentity match (VVNM) includes the same yellow can and red
box used in the ABLA, but the sample stimuli used are two three-letter words cut from
wood and thick enough to be described as three-dimensional. The three-dimensional
words bear no formal similarity to the yellow can and red box and so the relationship
between them is arbitrary and defined by the experimenter. All other procedures, including prompting strategies and passing or failing criteria, are the same as those used in other
levels of the ABLA. Sakko et al. (2004) found that the VVNM applied after level 4 and
before level 5 of the standard ABLA hierarchy. However, Jackson (2006) suggests that
the ability to make nonidentity or arbitrary visual discriminations may represent a higher
level of difficulty than ABLA level 6. Further replication is needed on a larger scale in
order to resolve this issue and include the VVNM as a level of the ABLA hierarchy.

The Practical Utility of the ABLA for Programming Learning

Given the widespread use of simple and conditional discriminations that are formed
both within and across modalities, it is clear that the ABLA can be used as an assessment
tool to make decisions about programming for learning. Failure on a particular educational, vocational, or daily living task can often be readily explained by the use of the
ABLA. For example, an individual who is unable to complete a sorting task may be found
to fail ABLA level 4, because it involves making visual-visual identity or quasi-identity
conditional discriminations. Alternatively, an individual who does not respond correctly
to verbal instructions after many training trials may be found to fail ABLA levels 5 and
6. It then becomes obvious that the ABLA not only provides a picture of that individuals
behavioral repertoire with respect to discrimination abilities involved in the task, but it
also suggests a course of action for remediation of such problems. This feature of the
ABLA in revealing repertoire limitations may be particularly relevant to descriptive and
experimental procedures for determining the function of challenging behaviors (Iwata,
Dorsey, Slifer, Bauman, & Richman, 1994). For example, if escape from task demands
is identified as the function of disruptive behavior, the ABLA may quickly demonstrate
that the individual cannot make auditory-visual discriminations and therefore cannot
comprehend spoken instructions. It then seems plausible to assume that an instructional
environment in which the individual does not understand the instructions, and has no
way to communicate this, may become unpleasant and the individual may attempt to
escape the situation. Although other functional assessment procedures will be useful in
such a situation, the ABLA will highlight such areas of difficulty and suggest a starting
point for programmed learning.
In the acquisition of new skills, an individuals difficulty learning one particular skill
may initially seem puzzling, but further analysis of the discriminations involved and
assessment of the individuals ABLA level may shed light on the problem. For example,
the failures of an individual who has experienced thousands of trials attempting to teach

a letter sound with a single letter printed on a flash card may initially seem strange;
however, the ABLA may reveal that the individual is not able to make auditory-visual
discriminations. This would clarify the nature of the problem, ruling out the possibility
that it could be a hearing or motivation problem and showing that it is actually a skill
building issue. Furthermore, the ABLA results will identify the individuals level of competence, and the next discrimination in the hierarchy that should be taught. In addition,
it may be appropriate for individuals to learn other skills that involve discriminations that
they are capable of, allowing them to be successful and acquire new skills within that
domain.
Other areas where the ABLA has general utility involve skills that seem comparable
in the natural environment but in actuality, when analyzed in view of ABLA level, are
not. For example, an individual may appear to be able to follow some simple spoken
instructions in certain situations but seem unable to respond to others. This may initially
seem like a problem of motivation, distraction, or a variety of other issues. However,
an analysis of ABLA level may reveal that the individual is only able to make visual
conditional discriminations and is in fact not able to make cross-modal auditory-visual
discriminations. Often further examination of the instructions to which the individual
responds correctly may allow the analyst to see that these are given in close proximity
to the individual, and with the addition of visual elements such as gestures and facial
expressions. When responding to these instructions the individual is able to use the visual
cues to make visual discriminations. The accuracy of these is of course a function of the
consistency of gestures and facial expressions across people and instances, although these
can be surprisingly consistent. The instructions to which the individual does not respond
correctly may be somewhat stripped of such visual features (for example, they are given
from a distance, the instructor is not looking at the individual, there is a lack of gestures
or facial expressions, or instructions are given using some type of speaker device where the
instructor is not physically present).
In all of these situations, the ABLA proves useful in at least four ways: First, it encourages an analysis of tasks with respect to the discriminations that are actually involved.
Second, it provides a measure of areas of competence with types of discriminations and
tasks involving these areas of competence. Third, it shows types of discriminations with
which the individual can be expected to have difficulty. Fourth, it provides direction for
future learning. Given that these discrimination types appear as a hierarchy of complexity, instructors should begin with the lowest level at which the individual was unsuccessful and begin teaching there. Although research suggests that learning in these areas
may take thousands of trials, it may be effective and worthwhile for that individual.
Furthermore, such efforts may lead to the discovery of additional techniques to accelerate
this learning.

The ABLA and Equivalence Relations

As we have seen, the ABLA is a practical tool that involves some of the building blocks
for the formation of more complex stimulus relationships. The different discriminations
within the hierarchy of the ABLA can be described as (1) a simple motor response, (2)
a simple position discrimination, (3) a simple visual discrimination, (4) a quasi-identity

conditional visual discrimination, (5) an auditory-position or auditory-visual conditional

discrimination, and (6) an auditory-visual conditional discrimination. (It is tempting to
consider level 1 a motor imitation response. For persons with that repertoire, that is probably an accurate description of the variables controlling their behavior. However, what
then would explain performance on the same difficulty of matching at level 4 if the
person fails that level? An alternative analysis is that the person without visual identity
matching skills passes level 1 due to the demonstration and repeated overcorrection for
error responses, together with the errorless feature of placing only one container on the
table. Both of these explanations could be tested empirically.)
Virtually all demonstrations of stimulus equivalence or other forms of derived relational responding involve conditional discriminations (see chapter 8 of this volume). The
original demonstrations of equivalence (Sidman, 1971; Sidman & Tailby, 1982) and the
great majority of studies that followed involved cross-modal (auditory-visual) conditional
discriminations. Indeed, a subsequent development was the controversy surrounding
the role of language ability for the formation of equivalence or other such derived relational responding (Hayes, 1989; Horne & Lowe, 1996, 1997; see also chapter 4 of this
volume).
Spoken language typically involves cross-modal equivalence relations emerging from
a history of cross-modal conditional discriminations. However, if language is not a necessary condition for such equivalence relations, it follows that equivalence may be possible within other sensory modalities. Although the literature on this subject is sparse,
within-sensorial-modality equivalence has been demonstrated for tactile-tactile relations
(Belanich & Fields, 1999, visual-visual relations (Green, 1990), and auditory-auditory
relations (Sidman, 1994). The fact that equivalence is possible within one sensory modality and may not be dependent upon language suggests that perhaps it is simply conditional discrimination performance that is necessary for the formation of derived stimulus
relations. With the necessity of language having been widely debated in the literature,
matching to sample is the most common training and testing technique in the stimulus
equivalence literature. The formation of conditional discriminations as the basis for a
variety of derived relations renders the ability to make such discriminations central to this
area. Although it is a straightforward conclusion, the relationship between equivalence
and conditional discriminations makes salient the findings of studies on the ABLA for
derived stimulus relations and language acquisition (Jackson, 2006; Jackson, Williams, &
Biesbrouck, 2006).

Future Research
The ABLA has prompted research that has produced consistent outcomes regarding a
hierarchy of discrimination skills, accurate prediction of performance on similar skills,
and a practical testing instrument for determining basic functioning levels. Nonetheless,
questions remain concerning the relationship among the discriminations represented in
the ABLA as well as the nature of discriminations beyond the ABLA level 6.
One feature of the ABLA that may require some clarification is the difference between
true visual-visual identity conditional discrimination performance and quasi-identity

p erformance. Most researchers have accepted this discrepancy in the ABLA level 4 materials and, for the most part, have reported research participants responding as identity
conditional discrimination formation. However, we have worked with participants who
can match black and white materials identical in size but for whom a change of size or
color without detailed fading has resulted in errors.
In a similar vein, refinement is needed in comparing the discriminative abilities of
different populations when these populations are presented with human-produced auditory cues as opposed to more simple pure tones for auditory-visual conditional discriminations and auditory-auditory conditional discriminations. This information is directly
relevant for procedures that involve computer presentation of tasks as opposed to tabletop
presentations with three-dimensional objects. Human speech as an auditory stimulus is
extremely complex compared to a simple auditory tone. In addition, it has been our experience that many lower-functioning individuals do not respond well or at all to computer
screen presentations of information; this is similar to the difficulty of some individuals
with responding to pictures of objects as opposed to actual objects (Dixon, 1981; Dixon
& Dixon, 1978).
Although some information is currently available due to the ABLA research on visual
and auditory discriminations, more research is needed on the relative positioning of
simple auditory skills, visual-visual nonidentity matching skills, auditory-auditory identity matching skills, and auditory-auditory nonidentity matching skills. Further research
on the role of other sensory modalities, such as tactile, proprioceptive, olfactory, and taste
discriminations, and their relative positioning, if any, with auditory and visual skills is
also in order. Such information would be relevant for the development of meaningful
relational responding repertoires as observed in the development of language, and many
social conventions involving those modalities.
Directly related to these issues is the need for clarification of the role of conditionality in any or all of these discriminations and any hierarchical ordering observed across
sensory channels. To date, the major feature of the ABLA outcomes has been the apparent necessity of ABLA level 4 skills in order to observe auditory-visual performances.
Whereas this has large practical implications for clinicians and educators, it also has generated a certain uneasiness in researchers and clinicians alike in its suggestion that basic
visual skills precede auditory skills. This possibility also appears to fly in the face of the
concept of dominance or at least preference of auditory discriminations over visual discriminations as seen in developed adult language repertoires.
Finally, expansion of the ABLA to include discrimination skills beyond ABLA level 6
would also be of practical value. Although the current ABLA is predictive of testability on
standard intelligence tools (Richards et al., 2002), there remains a huge gap that needs to
be filled with respect to the nature of responding and its possible hierarchical complexity
between ABLA level 6 and the repertoires associated with lower and then higher intelligence scores. The identification, for example, of milepost skills that could be demonstrated to correlate with existing standardized measures of intellectual functioning would
represent the first empirical behavioral tools for the assessment of cognitive development.
Because these skills may be describable by more complex arrangements such as those
involved in relational responding and the formation of more complex relational networks,
this volume represents the first efforts in such a direction.

References
Belanich, J., & Fields, L. (1999). Tactual equivalence class formation and tactual-to-visual
cross-modal transfer. Psychological Record, 49, 7591.
Dixon, L. (1981). A functional analysis of photo-object matching skills of severely retarded
adolescents. Journal of Applied Behavior Analysis, 14, 465478.
Dixon, M. H., & Dixon, L. S. (1978). The nature of standard control in childrens matching-to-sample. Journal of the Experimental Analysis of Behavior, 30, 205212.
Green, G. (1990). Differences in development of visual and auditory-visual equivalence
relations. American Journal on Mental Retardation, 95, 260270.
Hayes, S. C. (1989). Nonhumans have not yet shown stimulus equivalence. Journal of the
Experimental Analysis of Behavior, 51, 385392.
Horne, P. J., & Lowe, C. F. (1996). On the origins of naming and other symbolic behavior. Journal of the Experimental Analysis of Behavior, 65, 185241.
Horne, P. J., & Lowe, C. F. (1997). Toward a theory of verbal behavior. Journal of the
Experimental Analysis of Behavior, 68, 271296.
Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S. (1994).
Toward a functional analysis of self-injury. Journal of Applied Behavior Analysis, 27,
197209.
Jackson, M. (2006). An examination of the relations between ABLA performance, language
ability, and within modal stimulus equivalence. Masters thesis, University of Nevada,
Reno.
Jackson, M., Williams, W. L., & Biesbrouck, J. (2006). Conditional discrimination
ability, equivalence formation and mental retardation: Implications for development
in children with developmental disabilities. Journal of Speech Language Pathology and
Behavior Analysis, 1, 2742.
Kerr, N., Meyerson, L., & Flora, J. A. (1977). The measurement of motor, visual, and auditory discrimination skills [Monograph issue]. Rehabilitation Psychology, 24, 95112.
Marion, C., Vause, T., Harapiak, S., Martin, G. L., Yu, C. T., Sakko, G., et al. (2003).
The hierarchical relationship between several visual and auditory discriminations and
three verbal operants among individuals with developmental disabilities. Analysis of
Verbal Behavior, 19, 91105.
Martin, G. L., & Yu, C. T. (2000). Overview of research on the assessment of basic learning abilities test. Journal on Developmental Disabilities, 7, 1036.
Martin, G. L., Yu, C. T., & Vause, P. (2004). Assessment of basic learning abilities test:
Recent research and future directions. In W. D. Williams (Ed.), Developmental disabilities: Etiology, assessment, intervention and integration. Reno, NV: Context Press.
Meyerson, L. (1977). AVC behavior and attempts to modify it [Monograph issue].
Rehabilitation Psychology, 24, 11922.

Richards, D. F., Williams, W. L., & Follette, W. C. (2002). Two new empirically derived
reasons to use the Assessment of Basic Learning Abilities. American Journal on Mental
Retardation, 107, 329339.
Sakko, G., Martin, T. L., Vause, T., Martin, G. L., & Yu, C. T. (2004). A visual-visual
nonidentity matching assessment is a worthwhile addition to the Assessment of Basic
Learning Abilities test. American Journal on Mental Retardation, 109, 4552.
Sidman, M. (1971). Reading and auditory-visual equivalences. Journal of Speech and
Hearing Research, 14, 513.
Sidman, M. (1994). Equivalence relations and behavior: A research story. Boston: Authors
Cooperative.
Sidman, M., & Tailby, W. (1982). Conditional discrimination vs. matching to sample: An
expansion of the testing paradigm. Journal of the Experimental Analysis of Behavior,
37, 522.
Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts.
Sparrow, S., Balla, D., & Ceccetti, D. (1984). Vineland adaptive behavior scales. Circle
Pines, MN: American Guidance Service.
Stubbings, V., & Martin, G. L. (1995). The ABLA test for predicting performance of
developmentally disabled persons on prevocational training tasks. International
Journal of Practical Approaches to Disability, 19, 1217.
Ward, R., & Yu, D. C. T. (2000). Bridging the gap between visual and auditory discrimination learning in children with autism and severe developmental disabilities. Journal
on Developmental Disabilities, 7, 142155.