ERIC EJ931957: Infant Sign Training and Functional Analysis

2011·0.16 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2011, 44, 305–314 NUMBER2 (SUMMER2011)
INFANT SIGN TRAINING AND FUNCTIONAL ANALYSIS
MATTHEW P. NORMAND, MYCHAL A. MACHADO, KRISTIN M. HUSTYI, AND
ALLISON J. MORLEY
UNIVERSITYOFTHEPACIFIC
We taught manual signs to typically developing infants using a reversal design and caregiver-
nominatedstimuli.Wedeliveredthestimulionatime-basedscheduleduringbaseline.During
theintervention,weusedprogressivepromptingandreinforcement,describedbyThompsonet
al.(2004,2007),toestablishmands.Followingsigntraining,weconductedfunctionalanalyses
andverifiedthatthesignsfunctionedasmands.Theseresultsprovidepreliminaryvalidationfor
the verbal behavior functional analysis methodology and further evidence of the functional
independence ofverbaloperants.
Keywords: sign language, functional analysis, verbalbehavior
_______________________________________________________________________________
Anumberofstudiespublishedinrecentyears However, the degree to which sign language
report the success and potential benefits of taught under well-controlled conditions gener-
teaching sign language to infants (e.g., Good- alizes to other settings has received relatively
wyn&Acredolo,1993;Goodwyn,Acredolo,& little attention. Thompson et al. (2007) report-
Brown, 2000). However, in many cases this ed that signing generalized across experimenters
research has been limited by the absence of and settings, but they assessed only two
systematic teaching procedures and the lack of participants, and the generalization sessions
any direct observation of sign training. Recent- incorporatedthesameteachingproceduresused
ly,Thompsonandcolleaguesdescribedeffective in the initial training. In addition, the experi-
prompting and reinforcement procedures for menters did not identify the conditions under
teaching sign language to infants (Thompson, which the sign would be expected to occur.
Cotnoir-Bichelman, McKerchar, Tate, & Dan- According to Skinner’s (1957) taxonomy of
cho,2007;Thompson,McKerchar,&Dancho, language, the basic unit of a verbal behavior
2004). The acquisition of sign language was analysisistheverbaloperant,describedinterms
accompanied by reductions in problem behav- of the primary controlling variables over the
ior (e.g., crying and whining) when sign verbal response form, be it manual or vocal. In
training was combined with extinction for recent years, the clinical functional analysis
problem behavior (Thompson et al., 2007). methodology first described by Iwata, Dorsey,
These reports are important contributions to Slifer, Bauman, and Richman (1982/1994) has
the infant sign-language literature because of been modified for the purposes of evaluating
their technological precision and sound exper- the function of emerging language in young
imental methodology. children with autism and other developmental
disabilities according to Skinner’s functional
taxonomy (Kelley et al., 2007; LaFrance,
We thank Holly White, Jennifer Moreno, Rodnesha Wilder, Normand, & Squires, 2009; Lerman
Dunn, Dawn Howard, and the staff and management at et al., 2005; Normand, Severtson, & Beavers,
the Garden of Eden Day Care Center for their assistance
2008). This methodology shows promise for
inconducting thisstudy.
AddresscorrespondencetoMatthewNormand,Depart- assessingtheestablishedfunctionofsignstaught
mentofPsychology,UniversityofthePacific,3601Pacific to young children using general acquisition
Ave., Stockton, California 95211 (e-mail: mnormand@
strategies such as those described by Thompson
pacific.edu).
doi:10.1901/jaba.2011.44-305 et al. (2004, 2007). A more complete under-
305
306 MATTHEW P. NORMAND et al.
standing of the variables that evoke and Lerman et al., but they omitted the intraverbal
maintain signs would enable a better under- condition andmodified the echoiccondition so
standing of the conditions under which one as to be appropriate for a manual sign rather
would expect the sign to occur, and would than a vocal response.
suggest the operant functions to be targeted in The present study extends the existing
future teaching sessions if certain functional research on sign-language training in infants
analysis conditions did not evoke functional andyoungchildrenbyreplicatingtheprocedure
communication. described by Thompson et al. (2007) and
In their seminal report, Lerman et al. (2005) incorporating the trial-based verbal behavior
described a functional analysis in which a functional analysis methodology described by
number of test and control conditions corre- Kelley et al. (2007). Our goals were to test for
sponding to Skinner’s taxonomy were used to generalization in controlled situations and to
evaluatethefunctionofemergingspeechinfour assess the specific variables that controlled
children with developmental disabilities. The signing after structured sign training. To our
analysis was designed to test three of the knowledge, this is the first report of the verbal
behavior functional analysis methodology fol-
elementary verbal operants described by Skin-
lowing specific functional communication
ner: the mand, tact, and intraverbal. In
training with typically developing children,
addition, an echoic condition was introduced
and only the second to target manual rather
for one participant after undifferentiated re-
than vocal language.
sponding across the three primary test condi-
tions. Each condition involved the manipula-
tion of relevant antecedents and consequences, METHOD
and the dependent measure was the number of
Participants and Setting
target responses observed per minute. The
Three children participated. At the time of
results suggested that the verbal operants were
their initial participation, Julie and Ed were
functionally independent and that the func-
8 months old and Yvonne was 15 months old.
tional analysis distinguished among the various
None of the participants exhibited any vocal
functions. Moreover, the analysis most often
language, but Ed’s mother reported that he
identified a mand function for the target vocal
would occasionally use informal sign language
responses.
(i.e., communicative hand gestures not used in
Kelley et al. (2007) assessed the functions of American sign language [ASL]). Caregiver
vocal speech with four young children with nominations of preferred food and tangible
developmental disabilities using a trial-based items were used for the target response and
procedure in which the dependent measure reinforcer selection. All three participants
during each condition was the percentage of attended a local day-care center. Sessions for
trials during which the target response was Julie were scheduled during the center’s normal
observed. Despite the procedural modification, hours of operation. For Yvonne and Ed,
the results indicated at least one clear function sessions were scheduled in their homes at a
for each response assessed, similar to the results time mutually convenient for caregivers and
reportedbyLermanetal.(2005).However,the experimenters. To avoid contrived periods of
mand was not the most commonly identified food restriction or access, which might cause
verbaloperant(cf. LaFranceet al., 2009).More distress on the part of the infant, all sessions
recently, Normand et al. (2008) assessed the were scheduled at times that naturally followed
function of a manual sign with a young boy periods of food deprivation (e.g., first thing in
with autism using the procedure described by the morning, after nap time) or satiation (e.g.,
INFANT SIGN TRAINING 307
after a scheduled mealtime). All sessions were occurred during each interval. Interobserver
5 min in duration and were conducted in an agreement scores were scored for 89%, 73%,
unoccupiedareaofthehomeorday-carecenter. and 59% of sessions for Julie, Yvonne, and Ed,
Sessions were conducted one to three times per respectively. Mean agreement levels were 93%
day,severaldays perweek,andwere videotaped (range, 80% to 100%) for Julie, 89% (range,
for purposes of interobserver agreement and 67% to 100%) for Yvonne, and 89% (range,
intervention integrity analyses. The local insti- 60% to 100%) for Ed. For the functional
tutional review board approved all procedures. analysis sessions, the total number of trials on
which the two observers agreed on the occur-
Response Measurement and rence or nonoccurrence of an independent sign
Interobserver Agreement was divided by the total number of trials scored
During all sessions, observers recorded the and the quotient multiplied by 100%. Interob-
number of independent and prompted signs server agreement scores were scored for 100%,
exhibited by the participant. For all partici- 90%,and100%offunctionalanalysesforJulie,
pants, independent signing was defined as Yvonne, and Ed, respectively. Mean agreement
demonstrating the correct hand orientation levelswere100%forJulie,96%(range,90%to
and hand movement in the absence of any 100%) for Yvonne, and 99% (range, 96% to
prompts. A sign was scored as prompted if the 100%) for Ed.
response occurred following any visual (i.e.,
experimenter modeling of the sign, experiment- Design and Procedure
er showing the target item) or physical A reversal design was used to evaluate the
assistance (i.e., hand-over-hand guidance) from effectsofsigntrainingwithallparticipants.The
the experimenter. functional analysis conditions were arranged
The target response for Julie was an informal according to a multielement design.
sign for ‘‘applesauce,’’ defined as bringing the Initialbaseline.Thedesignatedreinforcerwas
open palm of one hand to the crown of the delivered on a fixed-time 10-s schedule
head. The target response for Ed was an (Thompson et al., 2004). Julie and Yvonne
informal sign for ‘‘rattle,’’ defined as the palm received a bite of applesauce or pears (respec-
of one hand hitting the thigh with a double tively) 10 s after consuming the previous bite.
movement. The target response for Yvonne was Ed received 10 s of access to the rattle.
a modified ASL sign (i.e., a communicative Reinforcer delivery was independent of the
handgestureapproximatingbutnotmatchinga participant’s behavior during these sessions.
formal ASL sign) for ‘‘pears,’’ defined as Sign training. After the initial baseline phase,
bringing the palm of one hand to the nose. sign training was implemented for each partic-
Interobserver agreement was calculated by ipant. For Julie and Yvonne, a model prompt
having a second independent observer collect was delivered immediately at the beginning of
data on independent and prompted signing each session and also according to a progressive
either during the session or from a videotape delay schedule after consumption of the
after the session. For sign training, each 5-min previously delivered reinforcer. If no signing
session was divided into 1-min intervals, and occurred within 5 s of the model prompt, the
the total number of intervals on which the two researcher physically prompted the target re-
observers agreed was divided by the total sponse. The delay to model prompts was
number of intervals scored and the dividend systematically increased across sessions on a
multiplied by 100%. An agreement was scored progressive schedule (e.g., 0 s, 5 s, 10 s, and
if both observers recorded independent or 20 s) until three consecutive sessions occurred
prompted signing as having occurred or not with 10% or more of signs occurring indepen-
308 MATTHEW P. NORMAND et al.
dently at the longest delay. The delay to the previously delivered reinforcer. Sessions
physical prompts remained constant at 5 s continued until an increase in independent
followingmodelprompts,andnopromptswere signs was observed. Booster sign-training ses-
delivered once the participant had met the sions were conducted on days between func-
response criterion at the longest delay. tional analysis sessions.
For Ed, a verbal prompt (‘‘What do you Functional analysis. Three test and corre-
want?’’) was delivered immediately at the sponding control conditions were arranged
beginning of each session and also according according to Skinner’s (1957) functional tax-
to the progressive delay schedule after removal onomy of verbal operants. Skinner described
of the previously delivered reinforcer. If no the response form of a particular verbal operant
signing occurred within 5 s of the verbal as occasioned by specific variables, either verbal
prompt, the experimenter repeated the verbal or nonverbal, and maintained by specific or
prompt while simultaneously holding the target generalized forms of reinforcement. The mand
item within Ed’s sight line. If no signing response form, for example, occurs when an
occurredwithin5softheverbal–visualprompt, organism is in a state of deprivation for a
the experimenter modeled the target response. specific reinforcer, and receipt of thatreinforcer
If no signing occurred within 5 s of the model contingent on the mand will strengthen similar
prompt, the experimenter physically prompted response forms in the future, under similar
Ed to perform the target response. The delay to conditions. The controlling variables for the
verbal prompts was systematically increased mand, tact, and echoic or mimetic (manual
across sessions (e.g., 0 s, 5 s, 10 s, and 20 s) equivalent to the echoic condition; Michael,
until three consecutive sessions occurred with 2004; Vargas, 1986) are outlined in Table 1,
10% or more of signs occurring independently alongwiththespecifictestandcontrolvariables
at the longest delay. The delay to verbal–visual, used in the functional analyses. Two test
model, and physical prompts remained con- sessions were conducted for every one control,
stant, and no prompts were delivered once the and the procedure was replicated twice for each
participanthadmettheresponsecriterionatthe of the three verbal operants tested. The specific
longest delay. arrangement of the conditions was similar to
The targetreinforcer foreach participant was that described by Lerman et al. (2005), but a
delivered immediately after all prompted and trial-based procedure was used (Kelley et al.,
independent signing. If the participant exhibit- 2007). However, unlike the Lerman et al. and
ed an approximation of the target sign, the Kelley et al. arrangements, the mimetic condi-
experimenterphysicallypromptedhimorherto tion was conducted for all participants. In
complete the target sign correctly and then addition,theintraverbalconditionwasomitted,
delivered the reinforcer. because the means by which one would select
Return to baseline. Procedures were identical the appropriate antecedent verbal response
to the initial baseline condition; however, the forms is unclear, making difficult any clear
reinforcer schedule was yoked to half the interpretation if responding is not observed in
interresponse time from the preceding sign- this condition.
training phase. All functional analysis conditions were ar-
Return to sign training (booster sessions). ranged to evaluate the function of the sign as a
Procedures were similar to the initial sign- mand, tact, or mimetic. For the mand test
training phase, except that the delay to each conditions, access to the target food item or
subsequent model or verbal prompt was held rattle was restricted for 60 min prior to the
constant at 5 s after consumption or removal of session,andaccesstothefooditemorrattlewas
INFANT SIGN TRAINING 309
Table 1
Summary ofControlling Variablesfor the Targeted Verbal Operants
Skinner’staxonomy Functionalanalysismethodology
Operant Antecedent Consequence Condition Antecedents Consequence
Mand Listenerplus Accesstothespecific Test 1hrdeprivation;listenerinclose Accesstothespecified
deprivationfora reinforcer proximity;verbalorvisual reinforcer
specificreinforcer prompt
Control Freeaccesstospecifiedreinforcer; Noprogrammed
listenerseatedacrossroom consequences
Tact Listenerplus Generalized Test Sightofitem;listenerinclose Briefpraise
nonverbal reinforcement proximity;verbalprompt
stimulus (e.g.,praise) Control Itemremoved;listenerseated Noprogrammed
acrossroom consequences
Echoic Listenerplus Generalized Test Itemremoved;listenerinclose Briefpraise
(mimetic) nonverbalstimulus reinforcement proximity;modelprompt
(e.g.,praise) Control Itemremoved;listenerseated Noprogrammed
acrossroom consequences
provided following each instance of the target independent signing (Kelley et al., 2007). The
sign. For example, Julie’s access to applesauce length of control sessions was yoked to the
was restricted for 60 min prior to conducting a length of the corresponding test condition. The
mand test session, and access to applesauce mand test conditions were 5 min in length (10
during the session was contingent on the trials), so the mand control conditions were
occurrence of the target sign. For the mand 5 min in length (10 30-s intervals).
control conditions, participants had free access
tothetargetfooditemorrattlefor30minprior
RESULTS
to the session (i.e., they were placed in close
proximity to the target item and were allowed Figure 1 shows the results of sign training,
toconsumeormanipulatetheitemfor30min), booster sessions, and functional analysis for
and there were no programmed consequences Julie. Julie did not sign during the initial
for independent signing. baseline phase. Her frequency of independent
Forthetacttestconditions,participants were signing reached the criterion to move to the
given free access to the target food items or delayed model prompt in four sessions
rattle for 30 min prior to and throughout the (20 min of training). Independent signing
test condition. Brief verbal praise was delivered increased steadily across 32 sign-training
following each instance of the target response. sessions (a total of 2.7 hr of sign training)
For the tact control, mimetic test, and mimetic and decreased to zero within five sessions after
control conditions, participants had free access reinstatement of baseline. When sign training
to the target food items or rattle for 30 min was resumed, independent signing increased to
prior to the session. The target food items or levels slightly higher than those observed
rattlewasnotpresentduringthesesessions.The during the initial sign-training phase within
experimenter provided no programmed conse- five sessions, and signing was maintained
quencesforindependentsigningduringthetact during three booster sessions conducted be-
and mimetic control conditions; independent tween functional analysis sessions. In total, the
signing in the mimetic test conditions resulted experimenter conducted 48 sessions (4 hr of
in brief verbal praise. training) with Julie.
Test conditions consisted of 10 trials, with Julie signed most often during the mand test
results expressed as the percentage of trials with sessions of the functional analysis, with a few
310 MATTHEW P. NORMAND et al.
Figure1. Julie’stotalnumberofindependentsignsfor
‘‘applesauce’’ in each session during baseline (BL), sign Figure2. Yvonne’stotalnumberofindependentsigns
training (ST), and booster training (Bst) (top). The for ‘‘pears’’ in each session during baseline (BL), sign
percentageoftrialsduringwhichJuliesigned‘‘applesauce’’ training (ST), and booster training (Bst) (top). The
during test and control sessions for each verbal operant percentage of trials during which Yvonne signed ‘‘pears’’
(bottom). during test and control sessions for each verbal operant
(bottom).
signs observed during the mimetic test sessions.
No signing was observed during the mand higher levels of independent signing than those
control, mimetic control, tact test, or tact observed in the initial sign-training sessions.
control sessions (Figure 1, bottom). These Four subsequent baseline sessions produced
results indicate that the sign for ‘‘applesauce’’ little to no signing. Sign training was again
functioned as a mand and as a mimetic. reinstated, and rates of independent signing
Figure 2 shows the results of sign training, comparable to the previous sign-training con-
booster sessions, and functional analysis for ditionwereobservedwithinthreesessions.Two
Yvonne. Yvonne did not sign during the initial booster sessions conducted on days between
baseline condition. Her frequency of indepen- functional analysis sessions resulted in slightly
dent signing reached the criterion to move to decreased rates of independent signing; howev-
the delayed model prompt in five sessions er, these levels still exceeded baseline levels. In
(25 min of training). Independent signing total, the experimenter conducted 38 sessions
increased steadily during 13 sign-training (3.2 hr of training) with Yvonne.
sessions (a total of 1.1 hr of sign training) and Yvonne signed often during all mand test
increased during the initial return to baseline; sessions, with a few signs observed during the
this was followed by an immediate decrease to mimetic test condition (Figure 2, bottom). No
near-zero levels. Reinstatement of sign training responding was observed in the mand control,
for 10 sessions (50 min of training) resulted in tact test, or tact control conditions. These
INFANT SIGN TRAINING 311
signed during the mand test and mimetic test
conditions and occasionally signed during the
tact test condition (Figure 3, bottom). Al-
though more variable than the other partici-
pants, these results suggest that the sign for
‘‘rattle’’functionedasamandandasamimetic.
DISCUSSION
All participants learned to sign under the
training conditions, with the frequency of
signing much higher in the sign-training
conditions than in baseline. Independent sign-
ing was observed within a few brief training
sessions for all participants. In addition, results
provided support for some functional indepen-
dence of the targeted verbal operants. That is,
the functional analysis identified a specific
condition or conditions that evoked signing.
Figure 3. Ed’s total number of independent signs for The greatest amount of signing was observed in
‘‘rattle’’ineachsessionduringbaseline(BL),signtraining the mand test condition, with signing also
(ST), and booster training (Bst) (top). The percentage of observed in the mimetic test condition. Signing
trials during which Ed signed ‘‘rattle’’ during test and
was almost never observed in the tact test
controlsessions for each verbal operant(bottom).
condition or in any of the control conditions.
These results are consistent with previous
results indicate that the sign for ‘‘pears’’ reports, in that young children (i.e., 8 to
functioned as a mand and as a mimetic. 15 months old) quickly acquired signing skills
Figure 3 shows the results of sign training, in a few training sessions (Thompson et al.,
booster sessions, and functional analysis for Ed. 2004, 2007) using a teaching procedure that
Ed signed independently during the initial consistedofdelayedmodelandphysicalprompts
baseline, but independent signing decreased to combinedwithreinforcement(Thompsonetal.,
zero within four sessions. His frequency of 2007). The functional analysis results support
independent signing reached the criterion to previousresearch,inthatthesignsoccurredonly
move to the delayed model prompt in four underspecifictestconditions(Kelleyetal.,2007;
sessions (20 min of training). Independent LaFrance et al., 2009; Lerman et al., 2005;
signing steadily increased during 19 sign- Normandetal.,2008),andrespondingoccurred
training sessions (a total of 95 min of training) mostofteninthemandcondition(Lermanetal.,
and then decreased to near zero within three 2005).However,unlikeinthesepreviousstudies,
sessions during the return to baseline. After the experimenters in the current study both
reinstatement of sign training, independent taught the manual signs and conducted the
signing increased to levels higher than those functionalanalysis,withsimilarsettingsusedfor
observed during the initial sign-training phase. training and testing. Thus, one might be more
Intotal,theexperimenterconducted32sessions likely to see clearly differentiated responding
(2.7 hr of training) with Ed. during the functional analysis, as opposed to
The results from Ed’s functional analyses are situationsinwhichthetargetverbalresponsewas
less clear than those of Julie and Yvonne. Ed establishedundersomewhatdifferentconditions
312 MATTHEW P. NORMAND et al.
and with different individuals than those experimental and test conditions arranged
involved in the functional analysis. That re- according to Skinner’s (1957) taxonomy, all
sponding also was observed in the mimetic test such results have been reported with young
condition,albeitlessfrequentlythaninthemand children with autism or other developmental
test condition, might be explained by the disabilities (Kelley et al., 2007; LaFrance et al.,
promptingproceduresusedduringsigntraining. 2009; Lerman et al., 2005; Normand et al.,
As a result of systematic model prompting, the 2008).
infants had the opportunity to sign under Given the potential generality of the meth-
conditions relevant both to mand and mimetic odology, it might be a useful tool for the
functions. These training procedures might longitudinal study of both typical and atypical
establish multiple verbal operants simultaneous- language development. For example, ongoing
ly, an outcome that could be beneficial in sign- descriptive analyses of child and caregiver
trainingprograms.Futureresearchisnecessaryto interactions could be conducted (e.g., Moerk,
evaluate this possibility. 1990;Vollmer,Borrero,Wright,VanCamp,&
These results extend previous research in Lalli, 2001) in combination with repeated
several ways. First, the present study served as a functional analyses conducted at regular inter-
preliminary validation of the verbal behavior vals. The results of these descriptive analyses of
functional analysis methodology in that the verbal behavior could be used to inform the
signs were taught under specific stimulus arrangement of functional analysis conditions.
conditions (i.e., as mands) and the functional Likewise, the results of the functional analysis
analysis subsequently identified a mand func- could inform the interpretation of possible
tionforthesigns.Thisisthefirstreportedstudy behavioral contingencies suggested by the
inwhichverbaloperantswereestablishedbythe results of the descriptive analysis. These types
experimentersandthenassessedusingtheverbal of arrangements could inform both basic and
behavior functional analysis methodology first applied research on language development.
describedbyLermanetal.(2005).AsNormand These areas are promising avenues for future
et al. (2008) noted, the functional analysis research. Similar methods that involve descrip-
methodology developed by Iwata et al. (1982/ tive and functional analyses of verbal behavior
1994) has been validated by a sizable research also could be used to inform clinical practice if
literature that has reported successful clinical results are used to track progress through
interventions based on the results of functional language acquisition programs and to identify
analyses. This kind of empirical validation is deficits in functional communication. Perhaps
not generally applicable to verbal behavior more important, a more complete understand-
analyses. A viable alternative to treatment ingofthetypesofstimuluscontrolproducedby
evaluations might be to establish novel verbal language acquisition procedures could lead to
operants under specific sources of stimulus more effective and efficient teaching technolo-
control and then assess function using the gies. Further research on verbal behavior
functional analysis methodology. The results functional analyses seems to be a promising
reported herein constitute such a validation. tool for such an endeavor.
Second, this is the first report of the verbal Third, the current study is one of only a few
behavior functional analysis methodology used empirical demonstrations of functional inde-
with typically developing individuals in general pendence among verbal operants exhibited by
andinfantsinparticular.Althoughanumberof typically developing individuals (cf. Lamarre &
studies have demonstrated that this methodol- Holland, 1985). The demonstration of func-
ogy produces differential responding across tionally independent verbal operants is difficult
INFANT SIGN TRAINING 313
in typically developing populations, because used in the present study (and in each of the
language acquisition occurs quickly and multi- verbal behavior functional analysis studies
ple sources of stimulus control are readily published to date) might be limited are that
established. Early sign training offers a unique (a) a nonverbal stimulus was absent in the tact
opportunity to capture early language develop- control condition, and (b) verbal discriminative
ment in typically developing infants and stimuli similar to those arranged in the test
toddlers and might also be applied to very early conditions were absent in the tact and mimetic
vocal development in this same population. A (echoic) control conditions. Future studies
number of inferences concerning language could address these concerns by incorporating
development can be made based on Skinner’s nonverbal stimuli that are irrelevant to the
(1957)verbalbehavior analysis.One exampleis target response into the tact control condition
the claim that mands, because they are the only and delivering irrelevant verbal prompts during
verbaloperantthatdirectlybenefitsthespeaker, the tact and mimetic control conditions on
developpriortoanyotheroftheverbaloperants schedules yoked to those in the test conditions
in early language learners (Bijou & Baer, 1965; (similar to the arrangement of the intraverbal
Lerman et al., 2005; Novak, 1996; Skinner, control condition reported by Lerman et al.,
1957; Sundberg, 2007; Sundberg & Michael, 2005).
2001). Although this is a plausible hypothesis, Future research could focus on several other
no strong empirical support for it exists in the areas. First, the sign-language acquisition pro-
literature. Functional analyses could be used to cedures reported thus far used delayed physical
evaluatesuchclaimsempiricallybyassessingthe ormodelpromptstoestablishsigning(Thomp-
variables that control the vocal or nonvocal son et al., 2004,2007).A procedure potentially
language observed in very young children as as or more effective might be to identify a
they begin to develop functional communica- salient motor response already exhibited by the
tion skills. infant and then to shape a clear manual sign.
In addition to the strengths noted above, Second, more attention should be given to the
several limitations of the current study warrant degree of generalization that results from the
attention. With respect to the sign-training reported sign-teaching procedures, especially as
procedures, the signs were taught only under assessed in natural settings with caregivers, and
mand conditions. Future research should eval- the variables that influence such generalization
uate the degree to which other verbal operants should be systematically investigated. Third, to
(e.g., tacts or intraverbals) could be established establish further the validity of the verbal
with infants and young toddlers. Also, the most behavior functional analysis methodology, an
relevant controlling variable (food deprivation) experimenter who is blind to the specific
was only loosely controlled during training. conditions under which the sign was taught
Although this seems not to have affected should conduct the functional analysis as a
acquisition adversely, more careful control over means of increasing its internal validity.
the relevant variables for each operant being In summary, the results reported herein
taught might improve acquisition rates or demonstrate that typically developing infants
generalization. and young toddlers can be taught to sign under
With respect to the verbal behavior func- conditions that correspond to those described
tional analysis, perhaps the most pressing by Skinner (1957) as controlling the mand and
question is the degree to which the control mimetic (echoic). In addition, the verbal
conditions were adequately arranged. Two behavior functional analysis methodology pro-
likely ways in which the control conditions duced differential signing across the test and
314 MATTHEW P. NORMAND et al.
control conditions, evoking the most signing in Lerman,D.C.,Parten,M.,Addison,L.R.,Vorndran,C.
M., Volkert, V. M., & Kodak, T. (2005). A
mandtestconditionsandsomewhatlesssigning
methodology for assessing the functions of emerging
in mimetic test conditions, thus providing one speech in children with developmental disabilities.
of the few demonstrations of functional inde- JournalofAppliedBehavior Analysis, 38,303–316.
Michael, J. L. (2004). Concepts and principles of behavior
pendence among verbal operants in typically
analysis (rev. ed.). Kalamazoo, MI: Society for the
developing children. The results of the func-
Advancement ofBehavior Analysis.
tional analysis also suggest that the methodol- Moerk,E.L.(1990).Three-term contingencypatternsin
ogy might be a useful tool for investigating mother-childinteractionsduringfirstlanguageacqui-
sition.JournaloftheExperimentalAnalysisofBehavior,
language development and for guiding lan-
54, 293–305.
guage-acquisition programs. Normand, M. P., Severtson, E. S., & Beaver, G. A.
(2008). A functional analysis of nonvocal verbal
behaviorofayoungchildwithautism.TheAnalysisof
REFERENCES VerbalBehavior,24,63–67.
Novak, G. (1996). Developmental psychology. Reno, NV:
Bijou, S. W., & Baer, D. M. (1965). Child development: ContextPress.
II. Universal stage of infancy. New York: Appleton- Skinner, B. F. (1957). Verbal behavior. New York:
Century-Crofts. Appleton-Century-Crofts.
Goodwyn, S. W., & Acredolo, L. P. (1993). Symbolic Sundberg, M. L. (2007). Verbal behavior. In J. O.
gestureversusword:Isthereamodalityadvantagefor Cooper, T. E. Heron, & W. L. Heward (Eds.),
onset of symbol use? Child Development, 64, Applied behavior analysis (2nd ed., pp. 526–547).
688–701. Upper SaddleRiver, NJ:Merrill/Prentice Hall.
Goodwyn, S. W., Acredolo, L. P., & Brown, C. (2000). Sundberg, M. L., & Michael, J. (2001). The benefits of
Impact of symbolic gesturing on early language Skinner’sanalysisofverbalbehaviorforchildrenwith
development. Journal of Nonverbal Behavior, 24, autism.BehaviorModification,25,698–724.
81–103. Thompson, R. H., Cotnoir-Bichelman, N. M.,
Iwata,B.A.,Dorsey,M.F.,Slifer,K.J.,Bauman,K.E., McKerchar, P. M., Tate, T. L., & Dancho, K. A.
& Richman, G. S. (1994). Toward a functional (2007). Enhancing early communication through
analysis of self-injury. Journal of Applied Behavior infant sign training. Journal of Applied Behavior
Analysis, 27, 197–209.(Reprinted from Analysis and Analysis, 40, 15–23.
Intervention in Developmental Disabilities, 2, 3–20, Thompson, R. H., McKerchar, P. M., & Dancho, K. A.
1982) (2004). The effects of delayed physical prompts and
Kelley, M. E., Shillingsburg, M. A., Castro, M. J., reinforcement on infant sign language acquisition.
Addison, L. R., LaRue, R. H., & Martins, M. P. JournalofAppliedBehavior Analysis, 37,379–383.
(2007).Assessmentofthefunctionsofvocalbehavior Vargas,E.A.(1986).Intraverbalbehavior.InP.N.Chase
in children with developmental disabilities: A repli- &L.J.Parrott(Eds.),Psychologicalaspectsoflanguage
cation. Journal of Applied Behavior Analysis, 40, (pp.128–151).Springfield, IL: Charles CThomas.
571–576. Vollmer,T.R.,Borrero,J.C.,Wright,C.S.,VanCamp,
LaFrance, D., Wilder, D. A., Normand, M. P., & C., & Lalli, J. S. (2001). Identifying possible
Squires, J. L. (2009). Extending the assessment of contingencies during descriptive analyses of severe
functions of vocalizations in children with limited behavior disorders. Journal of Applied Behavior
verbalrepertoires.TheAnalysisofVerbalBehavior,25, Analysis, 34, 269–287.
19–32.
Lamarre, J., & Holland, J. G. (1985). The functional Received May 7,2010
independence of mands and tacts. Journal of the Final acceptance September 20,2010
Experimental Analysisof Behavior, 43, 5–19. Action Editor,Rachel Thompson