-
525
Chapter 25 is devoted to verbal behavior, a distinguishing
feature of the humanbehavioral repertoire. Verbal behavior is at
once the phenomenon that makes humans particularly interesting and
the vehicle with which we express that inter-est. Verbal behavior
makes progress possible from one generation to the next and
enhances the development of sciences, technologies, and the arts.
Building onSkinner’s (1957) conceptual analysis, Mark Sundberg
presents verbal behavior in the context of typical human
development, with an emphasis on language as-sessment and
intervention programs for children with autism or other
develop-mental disabilities.
P A R T 1 0
Verbal Behavior
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 525
-
526
C H A P T E R 2 5
Verbal Behavior*
Key Termsaudienceautocliticautomatic punishmentautomatic
reinforcementconvergent multiple controlcopying a textdivergent
multiple controlechoicformal similarity
generic (tact) extensionsimpure
tactintraverballistenermandmetaphorical (tact) extensionmetonymical
(tact) extensionmultiple controlpoint-to-point correspondence
private eventssolistic (tact)
extensionspeakertacttextualtranscriptionverbal behaviorverbal
operant
Behavior Analyst Certification Board® BCBA® & BCABA®Behavior
Analyst Task List,© Third Edition
Content Area 3: Principles, Processes and Concepts.
3-15 Define and provide examples of echoics and imitation.
3-16 Define and provide examples of mands.
3-17 Define and provide examples of tacts.
3-18 Define and provide examples of intraverbals.
Content Area 9: Behavior Change Procedures
9-25 Use language acquisition programs that employ Skinner’s
analysis of verbalbehavior (i.e., echoics, mands, tacts,
intraverbals).
9-26 Use language acquisition and communication training
procedures.
© 2006 The Behavior Analyst Certification Board, Inc.,® (BACB®)
all rights reserved. A current version of thisdocument may be found
at www.bacb.com. Requests to reprint, copy, or distribute this
document and ques-tions about this document must be submitted
directly to the BACB.
*This chapter was written by Mark L. Sundberg.
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 526
-
Chapter 25 Verbal Behavior 527
Why should applied behavior analysts be con-cerned with verbal
behavior? A review of the de-finition of applied behavior analysis
as presented
in Chapter 1 can provide an answer to this question.
Applied behavior analysis is the science in which tacticsderived
from the principles of behavior are applied toimprove socially
significant behavior and experimenta-tion is used to identify the
variables responsible for be-havior change (p. 20).
Note the point to improve socially significant behavior.The most
socially significant aspects of human behaviorinvolve verbal
behavior. Language acquisition, social in-teraction, academics,
intelligence, understanding, think-ing, problem solving, knowledge,
perception, history,science, politics, and religion are all
directly relevant toverbal behavior. In addition, many human
problems, suchas autism, learning disabilities, illiteracy,
antisocial be-havior, marital conflicts, aggression, and wars,
involveverbal behavior. In short, verbal behavior plays a
centralrole in most of the major aspects of a person’s life, andin
the laws, conventions, archives, and activities of a so-ciety.
These topics are the main subject topics of mostintroductory
psychology textbooks. These are the sociallysignificant behaviors
that applied behavior analysts toaddress. However, the verbal
analysis of these topics hasjust begun, and a substantial amount of
work has yet tobe accomplished.
Verbal Behavior andProperties of LanguageForm and Function of
Language
It is important in the study of language to distinguish be-tween
the formal and functional properties of language(Skinner, 1957).
The formal properties involve the topog-raphy (i.e., form,
structure) of the verbal response,whereas the functional properties
involve the causes ofthe response. A complete account of language
must con-sider both of these elements.
The field of structural linguistics specializes in theformal
description of language. The topography of whatis said can be
measured by (a) phonemes: the individualspeech sounds that comprise
a word; (b) morphemes: theunits with an individual piece of
meaning; (c) lexicon:the total collection of words that make up a
given lan-guage; (d) syntax: the organization of words, phrases,
orclauses in sentences; (e) grammar: the adherence to es-tablished
conventions of a given language; and (f) se-mantics: what words
mean (Barry, 1998; Owens, 2001).
The formal description of a language can be accom-plished also
by classifying words as nouns, verbs, prepo-sitions, adjectives,
adverbs, pronouns, conjunctions, and 1For more detail see Mabry
(1994, 1995) and Novak (1994).
articles. Other aspects of a formal description of
languageinclude prepositional phrases, clauses, modifiers,
gerunds,tense markers, particles, and predicates. Sentences thenare
made up of the syntactical arrangement of the lexicalcategories of
speech with adherence to the grammaticalconventions of a given
verbal community. The formalproperties of language also include
articulation, prosody,intonation, pitch, and emphasis (Barry,
1998).
Language can be formally classified without the pres-ence of a
speaker or any knowledge about why thespeaker said what he did.
Sentences can be analyzed asgrammatical or ungrammatical from a
text or from a taperecorder. For example, incorrect use of word
tense canbe identified easily from a recording of a child
saying,“Juice all goned.”
A common misconception about Skinner’s analysisof verbal
behavior is that he rejected the formal classi-fications of
language. However, he did not find faultwith classifications or
descriptions of the response, butrather with the failure to account
for the “causes” orfunctions of the classifications. The analysis
of how andwhy one says words is typically relegated to the field
ofpsychology combined with linguistics; hence the field
ofpsycholinguistics.
Theories of Language
A wide variety of theories of language attempt to identifythe
causes of language. These theories can be classifiedinto three
separate, but often overlapping, views: biolog-ical, cognitive, and
environmental. The basic orientationof the biological theory is
that language is a function ofphysiological processes and
functions. Chomsky (1965),for example, maintained that language is
innate to hu-mans.1 That is, a human’s language abilities are
inher-ited and present at birth.
Perhaps the most widely accepted views of thecauses of language
are those derived from cognitive psy-chology (e.g., Bloom, 1970;
Piaget, 1952). Proponentsof the cognitive approach to language
propose that lan-guage is controlled by internal processing systems
that ac-cept, classify, code, encode, and store verbal
information.Spoken and written language is considered to be the
struc-ture of thought. Distinguishing between the biologicaland
cognitive views is often difficult; many are mixed(e.g., Pinker,
1994) and invoke cognitive metaphors suchas storage and processing
as explanations of languagebehaviors, or interchange the words
brain and mind (e.g.,Chomsky, 1965).
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 527
-
528 Part 10 Verbal Behavior
Developmentof Verbal BehaviorSkinner began working on a
behavioral analysis of lan-guage in 1934 as a result of a challenge
from Alfred NorthWhitehead,2 which Whitehead made when he was
seatednext to Skinner at a dinner at the Harvard Society of
Fel-lows. Skinner (1957) described the interaction as follows:
We dropped into a discussion of behaviorism which wasthen still
very much an “ism” and of which I was a zeal-ous devotee. Here was
an opportunity which I could notoverlook to strike a blow for the
cause . . . . White-head . . . agreed that science might be
successful in accounting for human behavior provided one made an
exception of verbal behavior. Here, he insisted some-thing else
must be at work. He brought the discussion to aclose with a
friendly challenge: “Let me see you,” he said,“account for my
behavior as I sit here saying ‘No blackscorpion is falling upon
this table.’” The next morning Idrew up the outline of the present
study. (p. 457)
It took Skinner 23 years to fill in the details of hisoutline,
which he published in his book Verbal Behavior(1957). The end
result was so significant to Skinner(1978) that he believed Verbal
Behavior would prove tobe his most important work. However,
Skinner’s use ofthe phrase prove to be 20 years after the book was
pub-lished indicated that his analysis of verbal behavior hadnot
yet had the impact that he thought it would.
There are several reasons for the slow appreciation ofVerbal
Behavior. Soon after the book was published, itwas met with
immediate challenges from the field of lin-guistics and the
emerging field of psycholinguistics. Mostnotably was a review by
Noam Chomsky (1959), a younglinguist from MIT who had published his
own account oflanguage (Chomsky, 1957) the same year Verbal
Behav-ior was published. Chomsky maintained that Skinner’sanalysis
was void of any value. Chomsky criticized everyaspect of the
analysis, but more so, he criticized the phi-losophy of behaviorism
in general. However, a reading ofChomsky’s review will reveal to
those who comprehendVerbal Behavior that Chomsky, like many
scholars,gravely misunderstood Skinner’s radical behaviorism,which
provided the philosophical and epistemologicalfoundations for
Verbal Behavior (Catania, 1972; Mac-Corqoudale, 1970).
Skinner never responded to Chomsky’s review, andmany felt this
lack of response was responsible for the
2Whitehead was perhaps the most prominent philosopher of the
time,known best for his landmark three-volume set coauthored with
BertrandRussell titled Principia Mathematica (1910, 1912,
1913).
widely held conclusion that Chomsky’s review was unan-swerable
and that Chomsky made valid criticisms.MacCorquodale (1970) pointed
out that the reason noone challenged Chomsky’s review was the
condescend-ing tone of the review, in addition to the clear
misunder-standings of Skinner’s behaviorism.
Skinner was not at all surprised by this reaction fromlinguists
because of their emphasis on the structure oflanguage rather than
its function. More recently however,a favorable review of Skinner’s
book from within the fieldof linguistics was published, recognizing
that Skinnerhas changed the history of linguistics (Andresen,
1991).
Although Skinner anticipated criticism from outsidethe field of
behavior analysis, he probably did not expectthe general
disinterest and often outspoken negative re-action to Verbal
Behavior from within the field. A num-ber of behaviorists have
examined this issue and havecollectively provided a list of reasons
behavior analystsdid not immediately embrace Verbal Behavior (e.g.,
Esh-leman, 1991; Michael, 1984; E. Vargas, 1986). Perhapsmost
troublesome to the behavior analysts of the timewas that Verbal
Behavior was speculative and did notcontain experimental data
(Salzinger, 1978).
The lack of research on verbal behavior continuedto concern
behavior analysts well into the 1980s (e.g.,McPherson, Bonem,
Green, & Osborne, 1984). However,this situation now appears to
be changing, and a numberof advances in research and applications
directly relate toVerbal Behavior (Eshleman, 2004; Sundberg,
1991,1998). Many of these advances are published in the jour-nal,
The Analysis of Verbal Behavior.
Defining Verbal BehaviorSkinner (1957) proposed that language is
learned behav-ior, and that it is acquired, extended, and
maintained bythe same types of environmental variables and
principlesthat control nonlanguage behavior (e.g., stimulus
con-trol, motivating operations, reinforcement, extinction).He
defined verbal behavior as behavior that is reinforcedthrough the
mediation of another person’s behavior. Forexample, the verbal
response “Open the door” can pro-duce the reinforcer of an open
door mediated through thebehavior of a listener. This reinforcer is
indirectly ob-tained, but is the same reinforcer that could be
obtainednonverbally by opening the door.
Skinner defined verbal behavior by the function ofthe response,
rather than by its form. Thus, any responseform can become verbal
based on Skinner’s functionaldefinition. For example, the early
differential crying of a2-month-old infant may be verbal, as would
other re-sponses such as pointing, clapping for attention,
gestures
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 528
-
Chapter 25 Verbal Behavior 529
such as waving one’s arm for attention, writing, or typ-ing. In
other words, verbal behavior involves a social in-teraction between
a speaker and a listener.
Speaker and Listener
The definition of verbal behavior makes a clear distinc-tion
between the behavior of the speaker and that of thelistener. Verbal
behavior involves social interactions be-tween speakers and
listeners, whereby speakers gain ac-cess to reinforcement and
control their environmentthrough the behavior of listeners. In
contrast with mostapproaches to language, Skinner’s verbal behavior
is pri-marily concerned with the behavior of the speaker. Heavoided
terms such as expressive language and receptivelanguage because of
the implication that these are merelydifferent manifestations of
the same underlying cogni-tive processes.
The listener must learn how to reinforce the speaker’sverbal
behavior, meaning that listeners are taught to re-spond to words
and interact with speakers. It is importantto teach a child to
react appropriately to the verbal stim-uli provided by speakers,
and to behave verbally as aspeaker. These are different functions,
however. In somecases learning one type of behavior (i.e., speaker
or lis-tener) facilitates learning another, but this must also
beunderstood in terms of motivating operations, antecedentstimuli,
responses, and consequences rather than in termsof learning the
meanings of words as a listener and thenusing the words in various
ways as a speaker.
Verbal Behavior: A Technical Term
In searching for what to call the subject matter of hisanalysis
of language, Skinner wanted a term that (a) em-phasized the
individual speaker, (b) referred to behaviorthat was selected and
maintained by consequences; and (c) was relatively unfamiliar in
the professions of speechand language. He selected the term verbal
behavior. How-ever, in recent years verbal behavior has acquired a
newmeaning, independent from Skinner’s usage. In the field of
speech pathology, verbal behavior has become synony-mous with
vocal behavior. Also, in psychology, the termnonverbal
communication, which became popular in the1970s, was contrasted
with the term verbal behavior, im-plying that verbal behavior was
vocal communication andnonverbal behavior was nonvocal
communication. Theterm verbal has also been contrasted with
quantitative asin GRE and SAT tests for college admissions. This
dis-tinction suggests that mathematical behavior is not
verbal.However, according to Skinner’s definition, much of
math-ematical behavior is verbal behavior. Noting that
verbalbehavior includes vocal-verbal behavior and nonvocal-verbal
behavior is sometimes confusing for those learn-ing to use
Skinner’s analysis.
Unit of Analysis
The unit of analysis of verbal behavior is the
functionalrelation between a type of responding and the same
in-dependent variables that control nonverbal behavior,namely (a)
motivating variables, (b) discriminative stim-uli, and (c)
consequences. Skinner (1957) referred to thisunit as a verbal
operant, with operant implying a typeor class of behavior as
distinct from a particular responseinstance; and he referred to a
set of such units of a par-ticular person as a verbal repertoire.
The verbal reper-toire can be contrasted with the units in
linguistics thatconsists of words, phrases, sentences, and the
meanlength of utterances.
Elementary VerbalOperantsSkinner (1957) identified six
elementary verbal oper-ants: mand, tact, echoic, intraverbal,
textual, and tran-scription. He also included audience relation and
copyinga text as separate relations, but in this discussion the
au-dience (or the listener) will be treated independently
andcopying a text will be considered a type of echoic be-havior.
Table 25.1 presents plain English descriptions of
Table 25.1 Plain English Definitions of Skinner’s Six Elementary
Verbal Operants
Mand Asking for reinforcers that you want. Saying shoe because
you want a shoe.
Tact Naming or identifying objects, actions, events, etc. Saying
shoe because you see a shoe.
Echoic Repeating what is heard. Saying shoe after someone else
says shoe.
Intraverbal Answering questions or having conversations in which
your words are controlled by other words. Saying shoewhen someone
else says, What do you wear on your feet?
Textual Reading written words. Saying shoe because you see the
written word shoe.
Transcription Writing and spelling words spoken to you. Writing
shoe because you hear shoe spoken.
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 529
-
530 Part 10 Verbal Behavior
these terms. Technical definitions and examples of
eachelementary verbal operant are provided in the
followingsections.
Mand
The mand is a type of verbal operant in which a speakerasks for
(or states, demands, implies, etc.) what he needsor wants. For
example, the behavior of asking for direc-tions when lost is a
mand. Skinner (1957) selected theterm mand for this type of verbal
relation because theterm is conveniently brief and is similar to
the plain Eng-lish words command, demand, and countermand. Themand
is a verbal operant for which the form of the re-sponse is under
the functional control of motivating op-erations (MOs) and specific
reinforcement (see Table25.2). For example, food deprivation will
(a) make foodeffective as reinforcement and (b) evoke behavior such
asthe mand “cookie” if this behavior has produced cookiesin the
past.
The specific reinforcement that strengthens a mandis directly
related to the relevant MO. For example, ifthere is an MO for
physical contact with one’s mother, thespecific reinforcement that
is established is physical con-tact. The response form may occur in
several topograph-ical variations such as crying, pushing a
sibling, reachingup, and saying “hug.” All of these behaviors could
bemands for physical contact if functional relations existamong the
MO, the response, and the specific reinforce-ment history. However,
the response form alone is insuf-ficient for the classification of
a mand, or any other verbaloperant. For example, crying could also
be a respondentbehavior if it were elicited by a conditioned or
uncondi-tioned stimulus.
Mands are very important for the early develop-ment of language
and for the day-to-day verbal inter-actions of children and adults.
Mands are the firstverbal operant acquired by a human child (Bijou
&Baer, 1965; Novak, 1996). These early mands usuallyoccur in
the form of differential crying when a child ishungry, tired, in
pain, cold, or afraid; or wants toys, at-tention, help, movement of
objects and people, direc-tions, or the removal of aversive
stimuli. Typicallydeveloping children soon learn to replace crying
withwords and signs or other standard forms of communi-cation.
Manding not only lets children control the de-livery of
reinforcers, but it begins to establish thespeaker and listener
roles that are essential for furtherverbal development.
Skinner (1957) pointed out that the mand is the onlytype of
verbal behavior that directly benefits the speaker,meaning that the
mand gets the speaker reinforcers suchas edibles, toys, attention,
or the removal of aversive stim-
uli. As a result, mands often become strong forms of ver-bal
behavior because of specific reinforcement, and thisreinforcement
often satisfies an immediate deprivationcondition or removes some
aversive stimulus. For exam-ple, young children often engage in a
very high rate ofmanding because of its effects on listeners. In
addition,much of the problem behaviors of children who haveweak,
delayed, or defective verbal repertoires may bemands (e.g., Carr
& Durand, 1985). Eventually, a childlearns to mand for verbal
information with who, what,and where questions, and the acquisition
of new verbalbehavior accelerates rapidly (Brown, Cazden, &
Bellugi,1969). Ultimately, mands become quite complex and playa
critical role in social interaction, conversation,
academicbehavior, employment, and virtually every aspect ofhuman
behavior.
Tact
The tact is a type of verbal operant in which a speakernames
things and actions that the speaker has directcontact with through
any of the sense modes. For ex-ample, a child saying “car” because
he sees a car is atact. Skinner (1957) selected the term tact
because itsuggests making contact with the physical environment.The
tact is a verbal operant under the functional con-trol of a
nonverbal discriminative stimulus, and it pro-duces generalized
conditioned reinforcement (see Table25.2). A nonverbal stimulus
becomes a discriminativestimulus (SD) with the process of
discrimination train-ing. For example, a shoe may not function as
an SD forthe verbal response “shoe” until after saying “shoe” inthe
presence of a shoe produces differential reinforce-ment.
A wide variety of nonverbal stimuli evoke tact re-lations. For
example, a cake produces nonverbal visual,tactile, olfactory, and
gustatory stimuli, any or all ofwhich can become SDs for the tact
“cake.” Nonverbalstimuli can be, for example, static (nouns),
transitory(verbs), relations between objects (prepositions),
prop-erties of objects (adjectives), or properties of
actions(adverbs); that is, nonverbal stimuli can be as simple asa
shoe, or as complex as a cancerous cell. A stimulusconfiguration
may have multiple nonverbal properties,and a response may be under
the control of those mul-tiple properties, as in the tact “The red
truck is on thelittle table.” Nonverbal stimuli may be observable
orunobservable (e.g., pain), subtle or salient (e.g., neonlights),
relational to other nonverbal stimuli (e.g., size),and so on. Given
the variation and ubiquity of nonver-bal stimuli, it is no surprise
that the tact is a primarytopic in the study of language.
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 530
-
Chapter 25 Verbal Behavior 531
Echoic
The echoic is a type of verbal operant that occurs whena speaker
repeats the verbal behavior of another speaker.For example, a child
saying “cookie” after hearing theword spoken by her mother is
echoic. Repeating thewords, phrases, and vocal behavior of others,
which iscommon in day-to-day discourse, is echoic also. Theechoic
operant is controlled by a verbal discriminativestimulus that has
point-to-point correspondence and for-mal similarity with the
response (Michael, 1982) (seeTable 25.2).
Point-to-point correspondence between the stim-ulus and the
response or response product occurs whenthe beginning, middle, and
end of the verbal stimulusmatches the beginning, middle, and end of
the response.Formal similarity occurs when the controlling
an-tecedent stimulus and the response or response product(a) share
the same sense mode (e.g., both stimulus andresponse are visual,
auditory, or tactile) and (b) physi-cally resemble each other
(Michael, 1982). In the echoicrelation the stimulus is auditory and
the response pro-duces an auditory product (echoing what one
hears), andthe stimulus and the response physically resemble
eachother.
Echoic behavior produces generalized conditionedreinforcement
such as praise and attention. The ability toecho the phonemes and
words of others is essential forlearning to identify objects and
actions. A parent mightsay, “That’s a bear, can you say bear?” If
the child can re-spond “bear,” then the parent says “Right!”
Eventually,the child learns to name a bear without the echoic
prompt.This often occurs in a few trials. For example, if a
childcan say “bear” (or a reasonable approximation) after aparent
says “bear,” then it becomes possible to teach thechild to say
“bear” in the presence of a picture of a bearor a bear at the zoo.
The echoic repertoire is very impor-tant for teaching language to
children with language de-lays, and it serves a critical role in
the process of teachingmore complex verbal skills (e.g., Lovaas,
1977; Sund-berg & Partington, 1998).
Motor imitations can have the same verbal prop-erties as echoic
behavior as demonstrated by their rolein the acquisition of sign
language by children who aredeaf. For example, a child may learn to
imitate the signfor cookie first, and then mand for cookie without
animitative prompt. Imitation is also critical for teachingsign
language to hearing children who are nonvocal.For the many children
who do not have an adequateechoic repertoire for vocal language
instruction, time isspent teaching echoic behavior rather than more
usefultypes of verbal behavior. A strong imitative
repertoirepermits a teacher to use sign language immediately to
instruct more advanced forms of language (e.g., mands,tacts, and
intraverbals). This allows a child to learnquickly to communicate
with others without using in-appropriate behavior (e.g., a tantrum)
to get what hewants.
Skinner also presented copying a text as a type ofverbal
behavior in which a written verbal stimulus haspoint-to-point
correspondence and formal similarity witha written verbal response.
Because this relation has thesame defining features as echoic and
imitation as it relatesto sign language, the three will be treated
as one cate-gory, echoic.
Intraverbal
The intraverbal is a type of verbal operant in which aspeaker
differentially responds to the verbal behavior ofothers. For
example, saying “the Buckeyes” as a resultof hearing someone else
say “Who won the game Satur-day?” is intraverbal behavior.
Typically developing chil-dren emit a high frequency of intraverbal
responses inthe form of singing songs, telling stories, describing
ac-tivities, and explaining problems. Intraverbal responsesare also
important components of many normal intellec-tual repertoires, such
as saying “Sacramento” as a resultof hearing “What is the capital
of California?”; saying“sixty-four” as a result of hearing “eight
times eight”; orsaying “antecedent, behavior, and consequence”
whenasked, “What is the three-term contingency?” The in-traverbal
repertoires of typical adult speakers includehundreds of thousands
of such relations.
The intraverbal operant occurs when a verbal dis-criminative
stimulus evokes a verbal response that doesnot have point-to-point
correspondence with the verbalstimulus (Skinner, 1957). That is,
the verbal stimulus andthe verbal response do not match each other,
as they doin the echoic and textual relations. Like all verbal
oper-ants except the mand, the intraverbal produces general-ized
conditioned reinforcement. For example, in theeducational context,
the reinforcement for correct an-swers usually involves some form
of generalized condi-tioned reinforcement such as “Right!” or
points, or theopportunity to move to the next problem or item
(seeTable 25.2).
An intraverbal repertoire facilitates the acquisitionof other
verbal and nonverbal behavior. Intraverbal be-havior prepares a
speaker to respond rapidly and accu-rately with respect to further
stimulation, and plays animportant role in continuing a
conversation. For exam-ple, a child hears an adult speaker say
“farm” in somecontext. If the stimulus farm evokes several relevant
in-traverbal responses, such as “barn,” “cow,” “rooster,”
or“horse,” then a child is better able to react to other parts
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 531
-
532 Part 10 Verbal Behavior
of an adult’s verbal behavior that may be related to a re-cent
trip to a farm. One might say that the child is nowthinking about
farms and now has relevant verbal re-sponses at strength for
further responses to the adult’sverbal behavior. An intraverbal
stimulus probes the lis-tener’s repertoire and gets it ready for
further stimulation.Collectively, mands, tacts, and intraverbals
contribute toa conversation in the following ways: (a) A mand
reper-toire allows a speaker to ask questions, (b) a tact
reper-toire permits verbal behavior about an object or eventthat is
actually present, and (c) an intraverbal repertoireallows a speaker
to answer questions and to talk about(and think about) objects and
events that are not physi-cally present.
Textual
Textual behavior (Skinner, 1957) is reading, without
anyimplications that the reader understands what is beingread.
Understanding what is read usually involves otherverbal and
nonverbal operants such as intraverbal behav-ior and receptive
language (e.g., following instructions,compliance). For example,
saying “shoe” upon seeing thewritten word shoe is textual behavior.
Understanding thatshoes go on a person’s feet is not textual.
Understandingis typically identified as reading comprehension.
Skinnerchose the term textual because the term reading refers
tomany processes at the same time.
The textual operant has point-to-point correspon-dence, but not
formal similarity, between the stimulusand the response product.
For example, (a) the verbalstimuli are visual or tactual (i.e., in
one modality) andthe response is auditory (i.e., another modality)
and (b) the auditory response matches the visual or tactualstimuli.
Table 25.2 presents a diagram of the textualrelation.
Textuals and echoics are similar in three respects:(a) They both
produce generalized conditioned rein-forcement, (b) both are
controlled by antecedent verbalstimuli, and (c) there is
point-to-point correspondencebetween the antecedent stimulus and
the response. Theimportant difference between textuals and echoics
isthat the response product of textual behavior (e.g., thespoken
word) is not similar to its controlling stimulus(e.g., the written
word evokes a spoken response or au-ditory response product). The
textual operant does nothave formal similarity, meaning that the
SDs are not inthe same sense mode and do not physically resemblethe
textual response. Words are visual and comprisedof individual
letters, whereas the reading response pro-duces an auditory
response product (which often iscovert) comprising phonemes. The
echoic responseproduct, however, does have formal similarity with
itscontrolling verbal stimulus.
Transcription
Transcription consists of writing and spelling words thatare
spoken (Skinner, 1957). Skinner also referred to thisbehavior as
taking dictation, with the key repertoires in-volving not only the
manual production of letters, butalso accurate spelling of the
spoken word. In technicalterms, transcription is a type of verbal
behavior in whicha spoken verbal stimulus controls a written,
typed, orfinger-spelled response. Like the textual operant, there
ispoint-to-point correspondence between the stimulus andthe
response product, but no formal similarity (see Table25.2). For
example, when asked to spell the spoken word“hat,” a response h-a-t
is a transcription. The stimulusand the response product have
point-to-point correspon-dence, but they are not in the same sense
mode or do notphysically resemble each other. Spelling English
words
Table 25.2 Antecedent and Consequent Controlling Variables for
Six Elementary Verbal Operants
Antecedent Variables Response Consequence
Motivating operations (4 hours without water) Mand (“water,
please”) Specific reinforcement (glass of water)
Nonverbal stimulus (see toy truck) Tact (“truck”) Generalized
conditioned reinforcement (GCSR) (praise and approval)
Verbal stimulus with point-to-point correspondence Echoic (say
“book”) GCSRand formal similarity (hear “book”)
Verbal stimulus without point-to point correspon- Intraverbal
(say “dogs”) GCSRdence or formal similarity (hear “cats and . .
.”)
Verbal stimulus with point-to-point correspondence, Textual (say
“apple”) GCSRwithout formal similarity (see apple written)
Verbal stimulus with point-to-point correspondence,
Transcription (write apple) GCSRwithout formal similarity (hear
“apple”)
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 532
-
Chapter 25 Verbal Behavior 533
is a difficult repertoire to acquire. Because many wordsin the
English language are not spelled the way theysound, shaping an
appropriate discriminative repertoire isoften difficult.
Role of the ListenerSkinner’s analysis of verbal behavior
focuses on thespeaker, whereas most linguists and psycholinguistic
ac-counts of language emphasize the listener. Skinner sug-gested
that the listener’s role is less significant thantypically assumed
because much of what is often de-scribed as listener behavior
(e.g., thinking, understanding)is more correctly classified as
speaker behavior. It is justthat often the speaker and listener
reside within the sameskin (as we will see in the following
discussion).
What role does the listener then play in Skinner’saccount of
language? In his analysis of listener behav-ior, Skinner pointed
out that a verbal episode requiresa speaker and a listener. The
listener not only plays acritical role as a mediator of
reinforcement for thespeaker’s behavior, but also becomes a
discriminativestimulus for the speaker’s behavior. In functioning
as adiscriminative stimulus, the listener is an audience forverbal
behavior. “An audience, then, is a discrimina-tive stimulus in the
presence of which verbal behavioris characteristically reinforced
and in the presence ofwhich, therefore, it is characteristically
strong” (Skin-ner, 1957, p. 172). When Skinner (1978) wrote
“verylittle of the behavior of the listener is worth
distin-guishing as verbal” (p. 122), he was referring to whenthe
listener serves as a discriminative stimulus in therole of an
audience.
A listener functions in additional roles, other than asa
mediator of reinforcement and a discriminative stim-ulus. For
example, verbal behavior functions as dis-criminative stimuli
(i.e., stimulus control) when aspeaker talks to a listener. The
question is, What are theeffects of verbal behavior on listener
behavior? A verbaldiscriminative stimulus may evoke echoic,
textual, tran-scription, or intraverbal operants of a listener. The
lis-tener becomes a speaker when this occurs. This isSkinner’s
point: The speaker and listener can and oftendo reside within the
same skin, meaning that a listenerbehaves simultaneously as a
speaker. The most signifi-cant and complex responses to verbal
stimuli occur whenthey evoke covert intraverbal behavior from a
listenerwho becomes a speaker and functions as her own audi-ence.
For example, a speaker’s verbal discriminativestimuli related to
Pavlov’s work on respondent condi-tioning, such as, “What was
Pavlov’s technique?” mayevoke a listener’s covert intraverbal
behavior such as
3In three of his books, Skinner devoted a full chapter to the
topic of think-ing: Science and Human Behavior (1953, Chapter 16),
Verbal Behavior(1957, Chapter 19), and About Behaviorism (1974,
Chapter 7) with severalsections dedicated to the topic of
understanding (e.g., Verbal Behavior,pp. 277–280; About
Behaviorism, pp. 141–142). A behavior analysis ofthinking and
understanding involves, in large part, situations in which
bothlistener and speaker reside within the same skin.
thinking, “He paired the sound of a metronome withmeat
powder.”3
Verbal stimulus control may also evoke a listener’snonverbal
behavior. For example, when someone says“Shut the door,” the
behavior of shutting a door is non-verbal, but shutting the door is
evoked by verbal stim-uli. Skinner (1957) identified this type of
listenerbehavior as understanding. “The listener can be said
tounderstand a speaker if he simply behaves in an appro-priate
fashion” (p. 277).
Verbal stimuli can become quite complex becauseseparating the
verbal and nonverbal behaviors of the lis-tener is difficult
(Parrott, 1984; Schoneberger, 1990,1991). For example, in following
a directive to buy a cer-tain type and style of pipe fitting at the
hardware store,success will involve both nonverbal behavior such as
dis-criminating among pipe fittings and verbal behavior suchas
self-echoic prompts (e.g., “I need a three-quarter-inchfitting,
three-quarter-inch”), tacts of the fittings (e.g.,“This looks like
three-quarter-inch”), and mands for in-formation (e.g., “Can you
tell me if this will fit a three-quarter-inch pipe?”).
Identifying Verbal OperantsThe same word (i.e., topography or
form of the behavior)can appear in the definitions of all of the
elementary ver-bal operants because controlling variables define
verbaloperants, not the form of the verbal stimuli. Verbal
be-havior is not classified or defined by its topography orform
(i.e., by the words themselves). The classification ofverbal
operants can be accomplished by asking a seriesof questions
regarding the relevant controlling variablesthat evoke a specific
response form (see Figure 25.1). Asample of verbal behavior
classification exercise is pre-sented in Table 25.3
1. Does an MO control the response form? If yes,then the operant
is at least part mand.
2. Does an SD control the response form? If yes, then:
3. Is the SD nonverbal? If yes, then the operant is atleast part
tact.
4. Is the SD verbal? If yes, then:
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 533
-
534 Part 10 Verbal Behavior
5. Is there point-to-point correspondence between theverbal SD
and the response? If not, then the operantis at least part
intraverbal. If there is point-to-pointcorrespondence, then:
6. Is there formal similarity between the verbal SD
and the response. If yes, then the operant must beechoic,
imitative, or copying a text). If not, then theoperant must be
textual or transcription.
Analyzing ComplexVerbal BehaviorAnalysis of more complex verbal
behavior includes au-tomatic reinforcement, tact extensions
(generalization),and private events. These topics are presented in
the fol-lowing sections.
Automatic Reinforcement
A common misconception about reinforcement in thecontext of
verbal behavior is that it occurs only when thelistener mediates
the reinforcer. When behavior occurswithout the apparent delivery
of reinforcement, it is oftenassumed that higher mental processes
are at work (e.g.,Brown, 1973; Neisser, 1976). Intermittent
reinforcementcan explain some behavior that occurs in the absence
ofan observed consequence, but not all such behavior. Somebehavior
is strengthened or weakened, not by externalconsequences, but by
its response products, which havereinforcing or punishing effects.
Skinner used the termsautomatic reinforcement and automatic
punishment ina number of his writings simply to indicate that an
ef-fective consequence can occur without someone provid-ing it
(cf., Vaughan & Michael, 1982).
Figure 25.3 Verbal Behavior Classification Chart.
Antecedent variables Verbal relation
UMO/CMO Yes Mand
NO
Nonverbal discrimination stimulus Yes Tact
No
Verbal discrimination stimulus Yes Point-to-point correspondence
No Intraverbal
Yes
Formal similarity Yes Echoic Imitation
Copy a text
No Textual Transcription
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 534
-
Table 25.3 Verbal Behavior Classification Exercises
As a result of . . . One has a tendency to . . . This is a(n) .
. .
1. seeing a dog say “dog” __________
2. hearing an airplane say “airplane” __________
3. wanting a drink say “water” __________
4. hearing “How are you?” say “I’m fine” __________
5. smelling cookies baking say “cookies” __________
6. tasting soup say “pass the salt” __________
7. hearing “book” write book __________
8. hearing “book” sign “book” __________
9. hearing “book” say “book” __________
10. hearing “book” say “read” __________
11. hearing “book” sign “read” __________
12. hearing “book” finger-spell “book” __________
13. seeing a book write book __________
14. wanting a book write book __________
15. signing “book” write book __________
16. hearing “color” say “red” __________
17. seeing a dog on the table say “get off” __________
18. seeing stop written hit the brakes __________
19. hearing “Skinner” write behavior __________
20. smelling smoke say “fire” __________
21. being hungry go to a store __________
22. seeing apple written sign “apple” __________
23. seeing 5 say “five” __________
24. wanting things say “thanks” __________
25. hearing “write your name” write your name __________
26. hearing “run” finger-spell “run” __________
27. seeing “home” signed sign “Battle Creek” __________
28. hearing a phone ring say “phone” __________
29. smelling a skunk say “skunk” __________
30. hearing “table” say “mesa” __________
31. being happy smile __________
32. hoping a pilot sees it writing SOS __________
33. wanting blue say “blue” __________
34. hearing “Red, white, and” say “blue” __________
35. tasting candy say “mmmm” __________
Provide examples of verbal behavior.
36. Give an example of a mand involving an adjective.
37. Give an example of a tact of a smell.
38. Give an example of a response that is part mand and part
tact.
39. Give an example of a response that is part tact and part
intraverbal.
40. Give an example of a tact involving multiple responses.
41. Give an example of an intraverbal using writing.
42. Give an example of receptive language using sign
language.Answers to verbal behavior classification exercises in
Table 25.3:1. T; 2. T; 3. M; 4. IV; 5. T/M; 6. M; 7. TX; 8. IV; 9.
E; 10. IV; 11. IV; 12. TR; 13. T; 14. M; 15. IV; 16. IV/T; 17. M;
18. NV; 19. IV; 20. T/M; 21. NV; 22. IV;23. IV; 24. M; 25. IV; 26.
TX; 27. IV; 28. T/M; 29. T; 30. IV; 31. NV; 32. M; 33. M; 34. IV;
35. T; 36. “I want the red one.”; 37. “Someone is smoking.”;38. “My
throat is dry.”; 39. “Macy’s.” When asked “Where did you buy
that?”; 40. “That’s a big burger!”; 41. A response to an e-mail;
42. Stoppingwhen someone signs “stop.”
535
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 535
-
536 Part 10 Verbal Behavior
Verbal behavior can produce automatic reinforce-ment, which has
a significant role in the acquisition andmaintenance of verbal
behavior. For example, automaticreinforcement may explain why a
typically developinginfant engages in extensive babbling without
the appar-ent delivery of reinforcement. Skinner (1957) pointed
outthat the exploratory vocal behavior of young childrencould
produce automatic reinforcement when those ex-ploratory sounds
match the speech sounds of parents,caregivers, and others.
Skinner (1957) described a two-stage conditioninghistory in
establishing vocal responses as automatic re-inforcers. First, a
neutral verbal stimulus is paired with anexisting form of
conditioned or unconditioned rein-forcement. For example, a
mother’s voice is paired withconditions such as presenting food and
warmth and re-moving aversive stimuli (e.g., medication on a
diaperrash). As a result, the mother’s voice, a previously neu-tral
stimulus, becomes a conditioned reinforcer. Themother’s voice will
now strengthen whatever behaviorprecedes it. Second, a child’s
vocal response as eitherrandom muscle movement of the vocal cords
or reflexivebehavior produces an auditory response that on
occasionmay sound somewhat like the mother’s words, intona-tions,
and vocal pitches. Thus, a vocal response can func-tion as
reinforcement by automatically increasing thefrequency of a child’s
vocal behavior.4
Automatic reinforcement also plays an important rolein the
development of more complex aspects of verbalbehavior, such as the
acquisition of syntax and gram-matical conventions. For example,
Donahoe and Palmer(1994) and Palmer (1996, 1998) suggested that a
child’suse of grammar produces automatic reinforcement whenit
sounds like the grammar used by others in the envi-ronment, but is
automatically punished when it soundsodd or unusual. Palmer (1996)
referred to this asachieving parity.
The stimulus conditions that evoke automatically re-inforced
behavior may be encountered everywhere be-cause each time a
response is automatically reinforced itmay alter the evocative
effect of any stimulus conditionthat might be present. For example,
a person may persistin singing or humming a theme song from a movie
whiledriving home from the movie because of the
two-stageconditioning process described previously. However,
thesong may be periodically evoked several hours, or evendays,
after the movie because each time the song is re-
4Miller and Dollard (1941) were perhaps the first to suggest
that a processlike automatic reinforcement might be partially
responsible for an infant’shigh rate of babbling. Since then, many
others have discussed and re-searched the role of automatic
reinforcement in language acquisition (e.g.,Bijou & Baer, 1965;
Braine, 1963; Miguel, Carr, & Michael, 2002; Mowrer,1950;
Novak, 1996; Osgood, 1953; Smith, Michael, & Sundberg,
1996;Spradlin, 1966; Staats & Staats, 1963; Sundberg, Michael,
Partington, &Sundberg, 1996; Vaughan & Michael, 1982; Yoon
& Bennett, 2000).
peated, a new stimulus such as a traffic light, a street
cor-ner, or a neon sign might acquire some degree of stimu-lus
control. The next time the person comes in contactwith, say, a red
light, there could be some tendency tosing or hum the song. This
effect might explain what isoften termed delayed echolalia observed
with childrenwho have autism (Sundberg & Partington, 1998). At
pre-sent, however, there have been no empirical investiga-tions of
the stimulus control involved in automaticconsequences, although it
certainly seems like an inter-esting and important area of
study.
Tact Extensions
Contingencies that establish stimulus and responseclasses and
generalization allow a variety of novel anddifferent discriminative
stimuli to evoke verbal behavior.Skinner (1957) said it this
way:
[A] verbal repertoire is not like a passenger list on a shipor
plane, in which one name corresponds to one personwith no one
omitted or named twice. Stimulus control isby no means so precise.
If a response is reinforced upona given occasion or class of
occasions, any feature ofthat occasion or common to that class
appears to gainsome measure of control. A novel stimulus
possessingone such feature may evoke a response. There are sev-eral
ways in which a novel stimulus may resemble astimulus previously
present when a response was rein-forced, and hence there are
several types of what wemay call “extended tacts” (p. 91).
Skinner (1957) distinguished four types of extendedtacts:
generic, metaphorical, metonymical, and solistic.The distinction is
based on the degree to which a novelstimulus shares the relevant or
irrelevant features of theoriginal stimulus.
Generic Extension
In generic extension, the novel stimulus shares all of
therelevant or defining features of the original stimulus.
Forexample, a speaker who learns to tact “car” in the pres-ence of
a white Pontiac Grand Am emits the tact “car” inthe presence of a
novel blue Mazda RX-7. A generic tactextension is evoked by simple
stimulus generalization.
Metaphorical Extension
In metaphorical extension, the novel stimulus sharessome but not
all of the relevant features associated withthe original stimulus.
For example, Romeo was experi-encing a beautiful, sunny, warm day,
and the exceptionalweather elicited respondent behaviors (e.g.,
good feel-ings). When Romeo saw Juliet, whose presence elicited
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 536
-
Chapter 25 Verbal Behavior 537
similar respondent behaviors like the sunny day, he said,“Juliet
is like the sun.” The sun and Juliet evoked a sim-ilar effect on
Romeo, controlling the metaphorical tactextension “Juliet is like
the sun.”
Metonymical Extension
Metonymical extensions are verbal responses to novelstimuli that
share none of the relevant features of the orig-inal stimulus
configuration, but some irrelevant but re-lated feature has
acquired stimulus control. Simply, oneword substitutes for another
in metonymical tact exten-sions, meaning that a part is used for a
whole. As exam-ples: Saying “car” when shown a picture of a garage,
orsaying “the White House requested” in place of “Presi-dent
Lincoln requested.”
Solistic Extension
Solistic extensions occur when a stimulus property thatis only
indirectly related to the tact relation evokes sub-standard verbal
behavior such as malaprops. For instance,using the solistic tact
extension, a person may say “Youread good” instead of “You read
well.” Saying “car” whenreferring to the driver of the car is a
solistic tact extension.
Private Events
In 1945 Skinner first described radical behaviorism,
hisphilosophy. At the core of radical behaviorism is theanalysis of
private stimuli (see also Skinner, 1953, 1974).Verbal behavior
under the control of private stimuli hasbeen a main topic of
theoretical and philosophical analy-ses of behavior ever since. In
1957 Skinner stated, “Asmall but important part of the universe is
enclosed withinthe skin of each individual. . . . It does not
follow that . . .it is any way unlike the world outside the skin or
insideanother’s skin” (p. 130).
A significant amount of day-to-day verbal behavioris controlled
in part by private events. What is commonlyreferred to as thinking
involves overt stimulus controland private events (e.g., covert
stimulus control). Theanalysis of private stimulation and how it
acquires stim-ulus control is complex because of two problems: (a)
Theparticipant can directly observe the private stimuli, butthe
applied behavior analyst cannot (a limiting factor inthe prediction
and control of behavior), and (b) privatestimulus control of verbal
episodes in the natural envi-ronment will likely remain private, no
matter how sensi-tive instruments will become in detecting private
stimuliand behaviors. Skinner (1957) identified four ways
thatcaregivers teach young persons to tact their private stim-
ulation: Public accompaniment, collateral responses,common
properties, and response reduction.
Public Accompaniment
Public accompaniment occurs when an observable stim-ulus
accompanies a private stimulus. For example, a fa-ther may observe
a child bump his head on a table whilechasing a ball. The public
stimuli are available to the fa-ther, but not the private and more
salient painful stimuliexperienced by the child. The father can
assume that thechild is experiencing pain because of his own
history ofbumping objects, and may say “Ouch,” or “You hurt
your-self.” In this way, the father is using the bump
(observablestimulus) as an opportunity to develop verbal
behaviorunder the stimulus control of a private stimulus. This
mayoccur with an echoic to the private event; later, the stim-ulus
control transfers to the private stimuli. Specifically,the child
may echo the father’s “ouch” while the painfulstimuli are present,
and quickly (depending on the child’shistory of echoic to tact
transfer) the painful stimuli aloneevokes the tact “ouch.”
Collateral Responses
Caregivers also teach young persons to tact their privatestimuli
by using collateral responses (i.e., observable be-havior) that
reliably occur with private stimuli. For ex-ample, the father may
not observe the child bump hishead, but may see the child holding
his head and crying.These collateral behaviors inform the father
that a painfulstimulus is present. The same training procedures
withpublic accompaniment can be used with collateral re-sponses.
Because the painful private stimuli are salient,only one trial may
be needed for the private stimuli toacquire stimulus control of the
tact relation.
Parents should use public accompaniment and col-lateral
responses during the beginning stages of tact train-ing. However,
even after developing a repertoire of tactingprivate events, a
parent or listener will have difficultyconfirming the actual
presence of the private event as in“My stomach hurts,” or “I have a
headache now.”
Also, learning to tact private behaviors is acquiredprobably
with public accompaniment and collateral re-sponses; for example,
private stimuli that evoke privateemotions (behaviors) that we tact
with words such ashappiness, sadness, fear, and being upset.
Learning totact such private events is difficult if the private
stimula-tion is not present during training. For example,
proce-dures that use pictures of people smiling and frowning(i.e.,
public stimuli) for teaching children to tact emo-tions will be
less effective than procedures that use vari-ables to evoke
pleasure or displeasure (i.e., privatestimuli) during training.
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 537
-
538 Part 10 Verbal Behavior
Common Properties
The two procedures described earlier use public stimulito
establish tacting of private events. Common proper-ties also
involve public stimuli, but in a different way. Aspeaker may learn
to tact temporal, geometrical, or de-scriptive properties of
objects and then generalize thosetact relations to private stimuli.
As Skinner (1957) noted,“[M]ost of the vocabulary of emotion is
metaphorical innature. When we describe internal states as
‘agitated,’‘depressed,’ or ‘ebullient,’ certain geometrical,
temporal,and intensive properties have produced a
metaphoricalextension of responses” (p. 132). Much of our verbal
be-havior regarding emotional events is acquired throughthis type
of stimulus generalization.
Response Reduction
Most speakers learn to tact features of their own bodiessuch as
movements and positions. The kinesthetic stim-uli arising from the
movement and positions can acquirecontrol over the verbal
responses. When movementsshrink in size (become covert), the
kinesthetic stimulimay remain sufficiently similar to those
resulting fromthe overt movements that the learner’s tact occurs as
aninstance of stimulus generalization. For example, a childcan
report imagining swimming, or can report self-talkabout a planned
conversation with someone, or can reportthinking of asking for a
new toy (Michael & Sundberg,2003). Responses produced by
private covert verbal be-havior can evoke other verbal behavior and
will be dis-cussed later in further detail.
Multiple ControlAll verbal behavior contains multiple functional
relationsamong antecedents, behavior, and consequences.
“[A]nysample of verbal behavior will be a function of manyvariables
operating at the same time” (Skinner, 1957, p.228). The functional
units of mands, tacts, echoics, in-traverbals, and textual
relations form the foundation of averbal behavior analysis. A
working knowledge of thesefunctional units is essential for
understanding the analy-sis of multiple control and complex verbal
behavior.
Convergent Multiple Control
Michael (2003) used the term convergent multiple con-trol to
identify when the occurrence of a single verbalresponse is a
function of more than one variable. The taskof an applied behavior
analyst is to identify the relevantsources that control an instance
of verbal behavior. Forexample, saying “Why did the United States
enter WorldWar II?” may be evoked by (a) MOs (making it part
mand), (b) verbal discriminative stimuli (making it partechoic,
intraverbal, or textual), (c) nonverbal stimuli(making it part
tact), or the presence of a specific audi-ence. For example, it’s
possible that an audience withcontempt for war (i.e., the MO)
evoked a mand, “Why didthe United States enter World War II?” The
question maybe more a function of this variable than an intraverbal
re-lated to the ongoing conversation or the nonverbal or tex-tual
stimuli that may be present in the room. On the otherhand, the
speaker may have no strong MO for the an-swer, but asks the
question because of its relation to MOsfor social reinforcement
related to political involvement.
Divergent Multiple Control
Multiple control also occurs when a single antecedentvariable
affects the strength of many responses. For ex-ample, a single word
(e.g., football) will evoke a varietyof intraverbal responses from
different people, and fromthe same person at different times.
Michael (2003) usedthe term divergent multiple control to identify
this typeof control. Divergent multiple control can occur also
withmand and tact relations. A single MO may strengthen avariety of
responses, such as food deprivation strength-ening the response
“I’m hungry,” or “Let’s go to a restau-rant.” A single nonverbal
stimulus can also strengthenseveral response forms, as when a
picture of a carstrengthens the responses “car,” “automobile,” or
“Ford.”
Thematic and Formal Verbal Operants
Skinner (1957) identified thematic and formal verbal op-erants
that function as sources of multiple control. Thethematic verbal
operants are mands, tacts, and intraver-bals and involve different
response topographies con-trolled by a common variable. For an
intraverbalexample, the SD “blue” can evoke verbal responses suchas
“lake,” “ocean,” and “sky.” The formal verbal oper-ants are echoic
(imitation, copying a text) and textual(and transcription) and are
controlled by a common vari-able, with point-to-point
correspondence. For example,the SD “ring” can evoke verbal
responses such as “sing,”“wing,” and “spring.”
Multiple Audiences
The role of the audience raises the issue of multiple
au-diences. Different audiences may evoke different re-sponse
forms. For example, two applied behavioranalysts talking (i.e., a
technical speaker and a technicalaudience) will likely use
different response forms thanwill a behavior analyst speaking with
a parent (i.e., a
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 538
-
Chapter 25 Verbal Behavior 539
technical speaker and a nontechnical audience). Apositive
audience has special effects, especially a largepositive audience
(e.g., as in a rally for a certain cause),as does a negative
audience. When the two audiences arecombined, the effects of the
negative audience are mostobvious: “[When a] seditious soapbox
orator sees a po-liceman approaching from a distance, his behavior
de-creases in strength as the negative audience becomesmore
important” (Skinner, 1957, p. 231).
Elaborating Multiple Control
Convergent multiple control occurs in most instances ofverbal
behavior. An audience is always a source of stim-ulus control
related to verbal behavior, even when aspeaker serves as his own
audience. In addition, it is alsothe case that more than one of the
controlling variables re-lated to the different verbal operants may
be relevant toa specific instance of verbal behavior. Convergent
controloften occurs with MOs and nonverbal stimuli, resultingin a
response that is part mand and part tact. For exam-ple, saying “You
look great” may be partly controlled bythe nonverbal stimuli in
front of a speaker (a tact), butalso by MOs related to wanting to
leave soon or wantingto avoid potential aversive events (a mand).
Skinner(1957) identified this particular blend of controlling
vari-ables as evoking impure tacts (i.e., impure because anMO
affects the tact relation).
Verbal and nonverbal stimuli can also share controlover a
particular response. For example, a tendency tosay “green car” may
be evoked by the verbal stimulus“What color is the car?” and the
nonverbal stimulus of thecolor green.
Multiple sources of control can be any combinationof thematic or
formal sources, even multiple sources fromwithin a single verbal
operant, such as multiple tacts ormultiple intraverbals. Skinner
pointed out that becausethese separate sources may be additive, the
“multiple cau-sation produces many interesting verbal effects,
includ-ing those of verbal play, wit, style, the devices of
poetry,formal distortions, slips, and many techniques of
verbalthinking” (pp. 228–229). Additional sources of controloften
reveal themselves—for example, when a speakerin the presence of an
obese friend who is wearing a newhat emits the “Freudian slip,” “I
like that fat on you.”Multiple sources of control often provide the
basis forverbal humor and listener enjoyment.
Autoclitic RelationThis chapter has emphasized that a speaker
can, and oftendoes, function as her own listener. The analysis of
howand why a speaker becomes a listener of her own verbalbehavior
and then manipulates her verbal behavior with
additional verbal behavior addresses the topic of the
auto-clitic relation. Skinner (1957) introduced the term
auto-clitic to identify when a speaker’s own verbal
behaviorfunctions as an SD or an MO for additional speaker ver-bal
behavior. In other words, the autoclitic is verbal be-havior about
a speaker’s own verbal behavior. Theconsequences for this behavior
involve differential rein-forcement from the ultimate listener,
meaning that thelistener discriminates whether to serve or not
serve as amediator of reinforcement for those verbal stimuli.
Aspeaker becomes a listener, and observer, of his own ver-bal
behavior and its controlling variables, and then in turnbecomes a
speaker again. This effect can be very rapidand typically occurs in
the emission of a single sentencecomposed of the two levels of
responses.
Primary and SecondaryVerbal Operants
Michael (1991, 1992) suggested that applied behavioranalysts
classify verbal behavior about a speaker’s ownverbal behavior as
primary (Level 1) and secondary(Level 2) verbal operants. In Level
1, MOs and/or SDsare present and affect the primary verbal operant.
Thespeaker has something to say. In Level 2, the speaker ob-serves
the primary controlling variables of her own ver-bal behavior and
her disposition to emit the primaryverbal behavior. The speaker
discriminates these con-trolling variables and describes them to
the listener. Asecondary verbal operant enables the listener’s
behavioras a mediator of reinforcement. For example, an MO orSD
evokes the response, “She is in Columbus, Ohio.” It is important
for listeners as reinforcement mediators todiscriminate the primary
variables controlling thespeaker’s behavior. The verbal operant
“She is in Colum-bus, Ohio,” does not inform the listener as to why
thespeaker said it. “I read in the Columbus Dispatch that sheis in
Columbus, Ohio,” informs the listener of the pri-mary controlling
variable. The first level is “She is inColumbus, Ohio” (the primary
verbal operant), and thesecond level is “I read in the Columbus
Dispatch” (theautoclitic).
Autoclitic Tact Relations
Some autoclitics inform the listener of the type of pri-mary
verbal operant the autoclitic accompanies (Peter-son, 1978). The
autoclitic tact informs the listener ofsome nonverbal aspect of the
primary verbal operant andis therefore controlled by nonverbal
stimuli. For example,a child’s statement, “I see Mommy” may contain
an au-toclitic tact. The primary verbal operant (i.e., the tact)
isthe nonverbal SD of (a) the child’s mother; (b) the re-sponse,
“Mommy”; and (c) the associated reinforcement
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 539
-
540 Part 10 Verbal Behavior
history. The secondary verbal operant (i.e., the autoclitictact)
is the speaker’s tact informing that a nonverbal SD
evoked the primary verbal operant. In this case, the non-verbal
SD was the visual stimulus of the child’s mother,and the response
“I see” informs the listener of the sourceof control that evoked
the primary tact. If the child heardhis mother, but did not see his
mother, the autoclitic tact“I hear” would be appropriate.
The listener may challenge the existence and natureof the
autoclitic tact—for example, “How do you knowit’s mommy?” The
challenge is one way that effectiveautoclitic behavior is shaped
and brought under appro-priate stimulus control.
Autoclitic tacts also inform the listener of the strengthof the
primary operant. In the examples of the verbal stim-uli, “I think
it is Mommy” and “I know it is Mommy,” “Ithink” informs the
listener that the source of control forthe primary tact “Mommy” is
weak; “I know” is strong.
Autoclitic Mand Relations
Speakers use autoclitic mands frequently to help the lis-tener
present effective reinforcers (Peterson, 1978). Aspecific MO
controls the autoclitic mand, and its role isto mand the listener
to react in some specific way to theprimary verbal operant. “I know
it is Mommy” may con-tain an autoclitic mand. For example, if “I
know” is nota tact of response strength, it may be an MO in the
samesense as “Hurry up.”
Autoclitic mands occur everywhere, but listenershave difficulty
recognizing the MO controlling the auto-clitic mand because the
sources of control are private.For example, hidden agendas as
autoclitic mands oftenrevel themselves only to the careful
observer. For exam-ple, a primary intraverbal such as an answer to
a questionabout the sale of a product may contain autoclitic
mands,as in “I’m sure you will be pleased with the sale,” inwhich
“I’m sure you will be pleased” is controlled by thesame MO that
might control the response “Don’t ask mefor any details about the
sale.”
Developing Autoclitic Relations
Speakers develop autoclitic relations in several ways.
Forexample, a father is wrapping a gift for his child’s mother,and
the child nearby says, “Mommy.” The father may askthe child to
identify the primary variables controlling theresponse by asking,
“Did you see her?” The father maydifferentially respond to “I see”
indicating that clearly“Mommy” is a tact and hide the gift; rather
than a mandfor “Mommy,” in which case, he keeps wrapping the
gift.The source of control for the response “Mommy” couldbe the
gift, as in, “That is for Mommy.” “That is for” (i.e.,the
autoclitic) informs the father that the gift is a nonver-
bal stimulus controlling the primary tact, “Mommy,” andthe
father continues wrapping the gift. As Skinner (1957)pointed out,
“[A]n autoclitic affects the listener by indi-cating either a
property of the speaker’s behavior or thecircumstances responsible
for that property” (p. 329).
Early language learners seldom emit autoclitic re-sponses.
Skinner was clear on this point: “In the absenceof any other verbal
behavior whatsoever autoclitics can-not occur. . . . It is only
when [the elementary] verbal op-erants have been established in
strength that the speakerfinds himself subject to the additional
contingencieswhich establish autoclitic behavior” (p. 330). Thus,
earlylanguage intervention program should not include auto-clitic
training.
Applications of Verbal BehaviorSkinner’s analysis of verbal
behavior provides a concep-tual framework of language that can be
quite beneficialfor applied behavior analysts. Viewing language
aslearned behavior involving a social interaction betweenspeakers
and listeners, with the verbal operants as thebasic units, changes
how clinicians and researchers ap-proach and ameliorate problems
related to language.Skinner’s theory of language has been
successfully ap-plied to an increasing number of human areas. For
ex-ample, the analysis has been used for typical languageand child
development (e.g., Bijou & Baer, 1965), ele-mentary and high
school education (e.g., Johnson &Layng, 1994), college
education (e.g., Chase, Johnson,& Sulzar-Azaroff, 1985),
literacy (e.g., Moxley, 1990),composition (e.g., J. Vargas, 1978),
memory (e.g., Palmer,1991), second language acquisition (e.g.,
Shimamune &Jitsumori, 1999), clinical interventions (e.g.,
Layng &Andronis, 1984), behavior problems (e.g., McGill,
1999),traumatic brain injury (e.g., Sundberg, San Juan, Dawdy,&
Arguelles, 1990), artificial intelligence (e.g., Stephens&
Hutchison, 1992), ape language acquisition (e.g., Sav-age-Rumbaugh,
1984), and behavioral pharmacology(e.g., Critchfield, 1993). The
most prolific application ofSkinner’s analysis of verbal behavior
has been to lan-guage assessment and intervention programs for
childrenwith autism or other developmental disabilities. This
areaof application will be presented in more detail in the
fol-lowing sections.
Language Assessment
Most standardized language assessments designed forchildren with
language delays seek to obtain an ageequivalent score by testing a
child’s receptive and expressive language abilities (e.g., Peabody
Picture
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 540
-
Chapter 25 Verbal Behavior 541
Vocabulary Test III [Dunn & Dunn, 1997], Comprehen-sive
Receptive and Expressive Vocabulary Test [Hammill& Newcomer,
1997]). Although this information is help-ful in many ways, the
tests do not distinguish among themand, tact, and intraverbal
repertoires, and important lan-guage deficits cannot be identified.
For example, thesetests assess language skills under the control of
discrim-inative stimuli (e.g., pictures, words, questions);
how-ever, a substantial percentage of verbal behavior is underthe
functional control of MOs. Manding is a dominatingtype of verbal
behavior, yet rarely is this repertoire as-sessed in standardized
testing. It is quite common to findchildren with autism or other
developmental disabilitieswho are unable to mand, but have
extensive tact and re-ceptive repertoires. If a language assessment
fails to iden-tify delayed or defective language skills that are
relatedto MO control, an appropriate intervention program maybe
difficult to establish. A similar problem is the failureto
adequately assess the intraverbal repertoire with moststandardized
assessments.
If a child with language delays is referred for a lan-guage
assessment, the behavior analyst should examinethe current
effectiveness of each verbal operant in addi-tion to obtaining a
standardized test from a speech andlanguage pathologist. The
behavior analyst would startby obtaining information about the
child’s mand reper-toire. When known motivating operations are at
strength,what behavior does the child engage in to obtain the
re-inforcer? When the reinforcer is provided, does the mandbehavior
cease? What is the frequency and complexityof the various mand
units? Information regarding thequality and strength of the echoic
repertoire can revealpotential problems in producing response
topographiesthat are essential for other verbal interactions. A
thor-ough examination of the tact repertoire will show the na-ture
and extent of nonverbal stimulus control over verbalresponses, and
a systematic examination of the receptiveand intraverbal
repertoires will show the control by ver-bal stimuli. Thus, a more
complete understanding of alanguage deficit, and hence a more
effective language in-tervention program, can be obtained by
determining thestrengths and weaknesses of each of the verbal
operants,as well as a number of other related skills (e.g.,
Parting-ton & Sundberg, 1998; Sundberg, 1983; Sundberg
&Partington, 1998).
Language Intervention
Skinner’s analysis suggests that a complete verbal reper-toire
is composed of each of the different elementary op-erants, and
separate speaker and listener repertoires. Theindividual verbal
operants are then seen as the bases forbuilding more advanced
language behavior. Therefore, alanguage intervention program may
need to firmly es-
tablish each of these repertoires before moving on to
morecomplex verbal relations such as autoclitics or
multiplycontrolled responses. Procedures for teaching mand,echoic,
tact, and intraverbal repertoires will be presentedbriefly in the
following sections, which will also include,some discussion of the
relevant research.
Mand Training
As previously stated, mands are very important to earlylanguage
learners. They allow a child to control the de-livery of
reinforcers when those reinforcers are most valu-able. As a result,
a parent or language trainer’s behavior(especially vocal behavior)
can be paired with the rein-forcer at the right time (i.e., when
the relevant MO foran item is strong). Mands also begin to
establish a child’srole as a speaker, rather than just a listener,
thus giving thechild some control of the social environment. If
mandsfail to develop in a typical manner, negative behaviorssuch as
tantrums, aggression, social withdrawal, or self-injury that serve
the mand function (and thereby controlthe social environment)
commonly emerge. Therefore, alanguage intervention program for a
nonverbal child mustinclude procedures to teach appropriate
manding. Theother types of verbal behavior should not be
neglected,but the mand allows a child to get what he wants, whenhe
wants it.
The most complicated aspect of mand training is thefact that the
response needs to be under the functionalcontrol of the relevant
MO. Therefore, mand training canonly occur when the relevant MO is
strong, and ulti-mately the response should be free from
additionalsources of control (e.g., nonverbal stimuli). Another
com-plication of mand training is that different response formsneed
to be established and brought under the control ofeach MO. Vocal
words are of course the most common re-sponse form, but sign
language, pictures, or written wordscan also be used.
The basic procedure for establishing mands consistsof using
prompting, fading, and differential reinforce-ment to transfer
control from stimulus variables to moti-vative variables (Sundberg
& Partington, 1998). Forexample, if a child demonstrates an MO
for watchingbubbles by reaching for the bottle of bubbles, then
smil-ing and laughing as he watches the bubbles in the air,
thetiming is probably right to conduct mand training. If thechild
can echo the word “bubbles” or an approximationsuch as “ba,”
teaching a mand can be easy (see Table25.4). The trainer should
first present the bottle of bub-bles (a nonverbal stimulus) along
with an echoic prompt(a verbal stimulus) and differentially
reinforce succes-sive approximations to “bubble” with blowing
bubbles(specific reinforcement). The next step is to fade theechoic
prompt to establish the response “bubble” under
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 541
-
542 Part 10 Verbal Behavior
Table 25.4 Teaching a Mand by Transferring Stimulus Controlto MO
Control
Antecedent Behavior Consequences
Motivating operation
Nonverbal stimulus “Bubbles” Blow bubbles
Echoic prompt
Motivating operation
Nonverbal stimulus “Bubbles” Blow bubbles
Motivating operation “Bubbles” Blow bubbles
the multiple control of the MO and the nonverbal stimu-lus (the
bottle of bubbles). The final step is to fade thenonverbal stimulus
to bring the response form under thesole control of the MO.
The easiest mands to teach in an early language in-tervention
program are usually mands for items for whichthe MO is frequently
strong for the child and satiation isslow to occur (e.g., food,
toys, videos). It is always im-portant to assess the current
strength of a supposed MOby using choice procedures, observation of
a child’s behavior in a free operant situation (i.e., no
demands),latency to contacting the reinforcer, immediate
con-sumption, and so forth. The goal of early mand trainingis to
establish several different mands by bringing dif-ferent response
forms (i.e., words) under the functionalcontrol of different MOs.
It is important to note that MOsvary in strength across time, and
the effects may be mo-mentary. In addition, the response
requirement placedon a child may weaken the strength of an MO,
makingmand training more difficult. Many additional strategiesexist
for teaching early mands to more difficult learners,such as
augmentative communication, physical prompts,verbal prompts, and
more careful fading and differentialreinforcement procedures (see
Sundberg & Partington,1998).
Manding continues to be an important part of a ver-bal
repertoire as other verbal operants are acquired. Soonafter mands
for edible and tangible reinforcers are ac-quired, a typical child
learns to mand for actions (verbs),attention, removal of aversive
stimuli, movement to cer-tain locations (prepositions), certain
properties of items(adjectives) and actions (adverbs), verbal
information(WH-questions), and so on. These mands are often
moredifficult to teach to a child with language delays becausethe
relevant MO often must be captured or contrived fortraining
purposes (Sundberg, 1993, 2004). Fortunately,Michael’s (1993)
classification of the different types ofMOs provides a useful guide
for capturing or contriv-ing MOs. For example, capturing a
transitive conditionedmotivative operation (CMO-T) in the natural
environ-
ment involves capitalizing on a situation in which onestimulus
increases the value of a second stimulus. A childwho likes fire
trucks may see a fire truck parked outsidethe window. This stimulus
condition increases the valueof a second stimulus condition, an
opened door, and willevoke behavior that has resulted in doors
opening in thepast. A skilled trainer would be watchful for these
eventsand would be quick to conduct a mand trial for the word“open”
or “out.” The work of Hart and Risley (1975)and their incidental
teaching model exemplify this teach-ing strategy.
Transitive CMOs can also be contrived to conductmand training
(e.g., Hall & Sundberg, 1987; Sigafoos,Doss, & Reichele,
1989; Sundberg, Loeb, Hale, & Eigen-heer, 2002). For example,
Hall and Sundberg (1987) useda contrived CMO-T procedure with a
deaf teenager withautism by presenting highly desired instant
coffee with-out hot water. The coffee altered the value of hot
waterand thereby evoked behavior that had been followed byhot water
in the past. During baseline this behavior con-sisted of tantrums.
Appropriate mands (i.e., signing “hotwater”) were easy to teach
when this CMO-T was in ef-fect by using the transfer of control
procedure describedearlier. In fact, a number of mands were taught
by usingthis procedure, and often the procedure led to the
emis-sion of untrained mands and a substantial reduction innegative
behavior.
Mand training should be a significant part of any in-tervention
program designed for children with autism orother severe language
delays. Without an appropriatemand repertoire, a child cannot
obtain reinforcementwhen MOs are strong, or have much control of
the so-cial environment. As a result, people who interact with
thechild may become conditioned aversive stimuli, and/orproblem
behaviors may be acquired that serve the mandfunction. These
behaviors and social relationships canbecome hard to change until
replacement mands are es-tablished. Teaching mands early in a
language interven-tion program may help to prevent the acquisition
ofnegative behaviors as mands. In addition, parents and
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 542
-
Chapter 25 Verbal Behavior 543
Table 25.5 Teaching Echoics by Using a Mand Frame
andTransferring Control from Multiple Control to Echoic Control
Antecedent Behavior Consequences
Motivating operation
Nonverbal prompt “Bubbles” Blow bubbles
Echoic stimulus
Nonverbal prompt
Echoic stimulus “Bubbles” Praise (GCR)
Echoic stimulus “Bubbles” Praise (GCR)
teachers are paired with successful manding and can be-come
conditioned reinforcers. If people become more re-inforcing to a
child, social withdrawal, escape andavoidance, and noncompliance
may be reduced.
Echoic Training
For an early language learner the ability to repeat wordswhen
asked to do so plays a major role in the develop-ment of other
verbal operants (as in the bubbles exampleearlier). If a child can
emit a word under echoic stimuluscontrol, then transfer of stimulus
control procedures canbe used to bring that same response form
under the con-trol of not only MOs, but also stimuli such as
objects(tacts) and questions (intraverbal). Because many chil-dren
with autism and other language delays are unableto emit echoic
behavior, special training procedures arerequired to develop the
echoic repertoire.
The first goal of echoic training is to teach the childto repeat
the words and phrases emitted by parents andteachers when asked to
do so. Once echoic control is ini-tially established, the goal
becomes to establish a gener-alized repertoire in which the child
can repeat novelwords and combinations. But the ultimate goal with
theechoic repertoire is to transfer the response form to
otherverbal operants. This transfer process can begin imme-diately
and is not dependent on the acquisition of a gen-eralized
repertoire. Several procedures will be describedto achieve the
first goal of establishing initial echoic stim-ulus control.
The most common form of echoic training is directechoic training
in which a vocal stimulus is presentedand successive approximations
to the target response aredifferentially reinforced. This procedure
involves a com-bination of prompting, fading, shaping, extinction,
andreinforcement techniques. Speech therapists commonlyuse prompts
such as pointing to the mouth, exaggeratedmovements, physical lip
prompting, and mirrors to watchlip movement. Successive
approximations to a target vo-calization are reinforced, and others
are ignored. Theprompts are then faded, and pure echoic responses
are
reinforced. For many children these procedures are ef-fective in
establishing and strengthening echoic controland improving
articulation. However, for some childrenthe procedures are
ineffective, and additional measuresare necessary.
Placing an echoic trial within a mand frame canoften be a more
effective procedure for establishingechoic stimulus control. The MO
is a powerful inde-pendent variable in language training and can be
tem-porally used to establish other verbal operants (e.g.,Carroll
& Hesse, 1987; Drash, High, & Tudor, 1999;Sundberg, 2004;
Sundberg & Partington, 1998). Forechoic training, an MO and
nonverbal stimulus can beadded to the target echoic antecedent as a
way to evokethe behavior (see Table 25.5). For example, if a
childdemonstrates a strong MO for bubbles, an echoic trialcan be
conducted while that MO is strong, and in thepresence of the
nonverbal stimulus of the bottle of bub-bles. These additional
sources of control can help toevoke the vocal response along with
the echoic prompt“Say bubbles.” The specific reinforcement of
blowingbubbles is then contingent on any successive approxi-mation
to “bubbles.” These additional antecedent vari-ables must be faded
out, and the reinforcement changedfrom the specific reinforcement
to generalized condi-tioned reinforcement. For some learners, the
transferfrom MO to echoic control may occur more quickly if
apicture of the object is used rather than the actual object(this
reduces the MO evocative effect).
Children with a low frequency of vocal behaviorsmay have
difficulty establishing echoic control. For thesechildren
procedures to simply increase any vocal behav-ior may facilitate
the ultimate establishment of echoiccontrol. One method is to
directly reinforce all vocal be-haviors. Taking this procedure one
step further, if a childrandomly emits a particular sound, the
behavior analystcan reinforce this behavior and conduct an echoic
trialwith that sound immediately after the delivery of
rein-forcement. Some children will repeat what they
initiallyemitted, and this interaction sets up some of the
basicvariables that may facilitate echoic control.
6216MC25.qxd_lpzIIII 10/20/06 1:26 PM Page 543
-
544 Part 10 Verbal Behavior
Table 25.6 Teaching Tacts by Using a Mand Frame and
Trans-ferring Control from Multiple Control to Nonverbal
Control
Antecedent Behavior Consequences
Motivating operation
Nonverbal prompt “Bubbles” Blow bubbles
Echoic stimulus
Nonverbal prompt
Echoic stimulus “Bubbles” Praise (GCR)
Nonverbal stimulus “Bubbles” Praise (GCR)
Automatic reinforcement procedures can also beused to increase
the frequency of vocal behavior. By pair-ing a neutral stimulus
with an established form of rein-forcement, the neutral stimulus
can become a conditionedreinforcer. For example, if just prior to
blowing bubblesthe trainer emits the word bubbles, bubbles can
becomea reinforcer. Research has shown that this pairing pro-cedure
can increase the rate of a child’s vocal play andresult in the
emission of targeted sounds and words thathad never occurred
echoically (Miguel, Carr, & Michael,2002; Sundberg, Michael,
Partington, & Sundberg, 1996;Smith, Michael, & Sundberg,
1996; Yoon & Bennett,2000). For example, Yoon and Bennett
(2000) demon-strated that this pairing procedure was more
successfulthan direct echoic training in producing targeted
sounds.Individual children who have difficulty acquiring anechoic
repertoire may benefit from this procedure or acombination of all
the procedures described in this section.
Tact Training
The tact repertoire is extensive and often the primary focusof
many language intervention programs. A child mustlearn to tact
objects, actions, properties of objects and ac-tions, prepositional
relations, abstractions, private events,and so on. The goal of the
teaching procedures is to bringa verbal response under nonverbal
stimulus control. If achild has a strong echoic repertoire, then
tact training canbe quite simple. A language trainer can present a
nonver-bal stimulus along with an echoic prompt, differentially
re-inforce a correct response, and then fade the echoicprompt.
However, for some children tact training is moredifficult, and
special procedures may be required.
A mand frame can also be used to establish tacting(Carroll &
Hesse, 1987). The procedure is similar to thatdescribed for
teaching an echoic response. Training be-gins with an MO for a
desired object, the nonverbal ob-ject, and an echoic prompt (see
Table 25.6). Using the
bubbles example, the first and second steps are the same,with
the goal being to free the response from motiva-tional control by
providing generalized conditioned re-inforcement rather than
specific reinforcement. As withechoic training, at this point in
the procedure the transfermay occur more quickly if a picture of
the object is usedrather than the actual object. Also, with some
children itmay be more effective to fade the echoic prompt
beforethe MO is faded. The third step in the procedure
involvesfading out the echoic prompt and bringing the responseunder
the sole control of the nonverbal stimulus, thus atact. Additional
nonechoic verbal prompts may also behelpful, such as “What is
that?,” but these too are verbalprompts that are additional sources
of control that needto be accounted for in the analysis of tact
acquisition(Sundberg & Partington, 1998).
Methods for teaching more complex tacts can alsomake use of the
transfer of stimulus control procedure.For example, teaching tacts
of actions requires that thenonverbal stimulus of movement be
present and a re-sponse such as “jump” be brought under the control
ofthe action of jumping. Teaching tacts involving prepo-sitions,
adjectives, pronouns, adverbs, and so on, alsoinvolves the
establishment of nonverbal stimulus con-trol. However, these
advanced tacts are often morecomplex than they appear, and
frequently the type ofstimulus control established in formal
training may notbe the same type of stimulus control that evokes
simi-lar tacts for typically developing children (Sundberg
&Michael, 2001). For example, some training programswith early
learner