Psychological Bulletin

1975, Vol. 82, No. 4, 565-580

Teaching Language to Nonverbal Children—With Emphasis

on Problems of Generalization
Sandra L. Harris
Rutgers—The State University
Operant conditioning techniques have been successfully used to teach nonverbal
children the skills necessary for functional language. A typical sequence of
training is: attention, nonverbal imitation, verbal imitation, and functional
speech. The basic components of these procedures have been well documented.
However, research showing means of facilitating the extension of functional
speech beyond the therapy room into the natural environment (i.e., generalizing
therapeutic gains) is limited in scope and adequacy. The heterogeneous popula-
tion included under the umbrella nonverbal makes research in this area espe-
cially difficult.

The past decade has witnessed the emer- these subjects have a full repertoire of re-
gence of powerful new operant conditioning sponses but have difficulty identifying the
procedures to establish speech in the reper- discriminative stimuli for speech (Isaacs,
toire of subjects previously consigned to Thomas, & Goldiamond, 1965; Sherman,
muteness. The present review summarizes the 1963, 1965).
operant research relevant to teaching lan- This review has been divided into four sec-
guage to nonverbal subjects; it also attempts tions: attention, nonverbal imitation, verbal
to identify problem areas in the research. In imitation, and functional language, which
particular it raises the question of the extent correspond to the stages of training typically
to which operantly trained speech generalizes used to teach nonverbal subjects.
to the natural environment. While much of
the cited literature is related to the theoretical ATTENTION
analysis of language acquisition, the present The first prerequisite for teaching speech is
discussion focuses more on the pragmatic a child who attends to his teacher. There are
issue of teaching speech rather than on theo- two basic settings for establishing attending
retical questions regarding learned versus behavior. In one, the child and adult are
innate language functions. seated facing each other in an open space. In
Few reviews are complete. The present the other, the child is enclosed in an isolation
article excludes unpublished dissertations as booth. Using the first procedure, Lovaas,
well as papers read at meetings. The term Berberich, Perloff, and Schaeffer (1966)
nonverbal has generally been limited to intel- seated their child and model with their heads
lectually retarded subjects who are mute or approximately 30 cm apart, and the adult
have minimal functional speech, although a used his legs to restrain the child from
few critical studies of somewhat more verbal leaving. The booth procedure was pioneered
subjects have been included. As a result, by Hewett (1965) and later used by Blake
studies of normally developing or "culturally and Moss (1967) and Fineman and Ferjo
deprived" children are excluded here, as are (1969). All researchers placed their subjects
those with "electively mute" subjects, since in booths in order to maximize the probability
that they would be facing the experimenter
This work was supported in part by a grant from
the Charles and Johanna Busch Memorial Fund of when the stimuli were presented. The direct
Rutgers, The State University of New Jersey. Peter contact procedure between experimenter and
E. Nathan made helpful suggestions on the manu- subject is more popular than the booth, prob-
script. ably because of simplicity, as well as the con-
Requests for reprints should be sent to Sandra L.
Harris, Psychology Department, Douglass College, cern that the booth creates specific stimulus
Rutgers—The State University of New Jersey, New conditions that might hamper later generali-
Brunswick, New Jersey 08903. zation. While Hewett reported no such dif-
565
566 SANDRA L. HARRIS

ficulty, the direct contact method still appears McConnelPs (1967) study suggests the
intuitively more parsimonious. extent to which eye contact may be under
The general pattern in establishing eye stimulus control. Having established eye con-
contact is to: (a) reinforce spontaneous eye tact with smiles and praise, he went on to
contact if it occurs or prompt eye contact show that the nature of specific toys in the
with food held near the model's face and the room influenced the degree of eye contact
verbal command, "Look at me" and then emitted by the subject. His data point to the
(b) reinforce on a continuous basis responses pervasive problem of teaching generalization
occurring within 5 seconds of the command. of response to the child with autistic behavior.
The reward schedule is gradually reduced to
an intermittent one (e.g., Brooks, Morrow, &
NONVERBAL IMITATION
Gray, 1968; Kozloff, 1973). The primary
variations around this pattern depend upon Once attending behaviors are established,
whether: (a) the subject is given a command, the next step in language training is teaching
(b) training is done individually or in a nonverbal imitative behaviors (e.g., Bricker &
group, (c) the eyes or the mouth are the Bricker, 1970; Buddenhagen, 1971; Kozloff,
focus of gaze behavior, and (d) primary or 1973), although not all researchers report this
secondary reinforcement is used. phase (e.g., Lovaas et al., 1966). Typically
Lovaas et al. (1966) rewarded their sub- the child is taught a series of gross motor
jects with food for visually fixing, without imitations such as clapping, standing, or,
command, upon the model's mouth to a cri- touching his toes. These gross movements
terion of 50% of each session. Hewett then progress to more refined movements
(1965) and Martin, England, Kaprowy, around the mouth (e.g., Marshall & Hegrenes,
Kilgour, and Pilek (1968) likewise waited for 1970; Slbane, Johnston, & Harris, 1968;
the child to attend and then rewarded the Stark, Giddan, & Meisel, 1968). Nonverbal
behavior. This strategy appears well suited imitation training may be done individually
in the group setting employed by Martin et or in a group (e.g., Borus, Greenfield, Spiegel,
al. While Gardner, Pearson, Bercovici, and & Daniels, 1973; Koegel & Rincover, 1974).
Bricker (1968) relied exclusively upon praise Interestingly there are few data document-
and physical contact to increase attending ing the extent to which nonverbal training
behavior in their subject, food is generally facilitates later verbal imitative training nor
heavily used in the early stages of training. the quantity of training necessary to enhance
Although most researchers describe rapid es- generalizations from nonverbal to verbal imi-
tablishment of eye contact, Kent, Klein, Falk, tation. The value of nonverbal imitation
and Guenther (1972) reported difficulty with training has been more a clinical assumption
their subjects. than an empirical fact. When Garcia, Baer,
Less formal strategies for establishing at- and Firestone (1971), using four retarded
tention include holding the child on one's lap subjects in a multiple-baseline design, ex-
(Kerr, Meyerson, & Michael, 1965) and using plored the degree of generalization across four
a mirror and glasses to induce the subject to topographically different classes of imitative
look at the experimenter (Colligan & Bellamy, behavior (small motor, large motor, short
1968). Ney (1973) examined the effects of vocal, and long vocal) they found a lack of
noncontingent, as contrasted with contingent, generalization across classes. Probe responses
reward upon eye contact. His observation of were imitated only when they were of the
increased eye contact during noncontingent type previously trained or currently being
reward, as compared to baseline, can best be trained. Buddenhagen (1971) likewise re-
explained as an experimental artifact (Harris, ported that the transition from nonverbal to
1974). Risley (1968), in a widely cited study verbal imitation was extremely difficult.
of the use of shock to eliminate dangerous In a related vein, Churchill, Hingtgen, and
climbing behavior, reported increased sitting their colleagues looked at the generalization
and eye contact as positive side effects of the of various forms of imitative behavior
aversive conditioning procedures. (Churchill, 1969; Hingtgen & Churchill,
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 567

1970; Hingtgen, Coulter, & Churchill, 1967). reinforcement and must generate novel re-
Their results point to the variability existing sponses that may or may not be reinforced.
among subjects and the need for caution in These requirements make language training
assuming that skills acquired in one modality difficult and enhance the importance of that
will generalize to another. Some subjects ac- literature which has dealt with the question
quired perceptual skills in 10 hours that other of generalized imitation, both verbal and non-
subjects had not acquired after 600 hours verbal. The identification of variables central
(Hingtgen & Churchill, 1969). to this process would certainly aid in the pro-
In an early study of nonverbal imitation grammed development of generalized verbal
in autistic children, Metz (1965) established imitation.
generalized imitation of nonverbal behavior An important study by Baer et al. (1967)
using praise and food as reinforcement. How- systematically evaluated the learning of gen-
ever, as Baer, Peterson, and Sherman (1967) eralized imitation in three retarded youngsters
pointed out, verbal praise was also delivered showing no preliminary evidence of imitation.
for new imitations, thus making it difficult to The authors trained a variety of imitative
separate the effects of generalization from the behaviors with shaping, fading, and physically
reinforcing properties of praise. This study moving the subject's body through the act.
also failed to include a reversal or a similar They used probes to test for generalization,
experimental manipulation, making it impos- a differential-reinforcement-of-other phase in
sible to attribute the increase in imitative which activities other than imitation were re-
behavior to training. Nevertheless, the dem- inforced to demonstrate the importance of rein-
onstration that imitative behavior could be forcement in maintaining imitation, and new
trained in subjects with little initial inclina- experimenters to show generalization from the
tion to imitate provided an impetus for original trainer to others. Generalized imita-
further research. tive behavior was established in all three sub-
In a study focused upon using imitation jects, who learned new responses in shorter
to build complex behaviors in schizophrenic periods of time and became increasingly likely
children, Lovaas, Freitas, Nelson, and Whalen to imitate new responses upon their initial
(1967) successfully established such activities introduction. In explaining their results Baer
as drawing, printing, self-care skills, and pre- et al. suggested that generalized imitation
school games by these means. The degree of occurs because similarity to the model takes
generalization and the use of shaping to on a reinforcing function as well as a discrim-
develop simple imitative acts into complex inative one. Similarity in and of itself could
skills that were maintained without immedi- then serve to reinforce imitative behavior.
ate external reinforcement suggests how useful Peterson (1968) pursued the studies of Baer
generalized imitative behavior can be in et al. (1967) using one of the subjects from
training nonverbal behavior. Nevertheless, as that earlier research. In his first experiment,
Lovaas et al. (1967) indicated, even after a he attempted to free one of the previously
year of intensive training there were limits learned imitative responses from the general
to the extent of generalized imitation ob- response class of imitative behavior. Although
served. The gap between the imitative it was possible to extinguish a single response
behavior of their subjects to the wide- under conditions of massed evocation without
ranging imitation of the normal child remains reinforcement, this same response continued
formidable. to be emitted when the evoking stimulus was
The maintenance of nonreinforced imita- interspersed among a series of imitative stim-
tion is a critical issue in the generalization of uli. Thus, the discrimination of this non-
appropriate verbal behavior. The goal of reinforced response proved difficult to estab-
language training is verbal behavior that is lish. Investigating the idea that similarity
maintained on the intermittent basis that between model and subject might be a fac-
characterizes reinforcement in the natural en- tor reinforcing continued imitative behavior
vironment. The normal child must emit famil- during extinction, Peterson hypothesized 'that
iar responses again and again with minimal if similarity per se were reinforcing, then non-
568 SANDRA L. HARRIS

imitative behavior, unlike imitative behavior, easier for the retarded nonverbal child to
Would extinguish under both massed and discriminate reinforced and nonreinforced be-
interspersed presentations. Contrary to pre- havior. If, for example, nonreinforced imita-
diction, the nonimitative responses resembled tions were always done in a dim light or were
imitative responses under both massed and accompanied by a buzzer, this cue might
interspersed conditions, suggesting that simi- reduce task complexity. Peterson's (1968)
larity per se was not the crucial variable massed extinction procedure might have
maintaining responding under interspersed served primarily to aid the child in her dis-
conditions. In his discussion, Peterson re- crimination of the nonreinforced task. The
ferred to Bandura's (1968) notion that diffi- observation that the extinguished response re-
culty in discrimination may account for the occurred when interspersed among imitative
maintenance of nonreinforced imitative be- stimuli might be interpreted to mean that the
havior. Massed evocation makes this dis- discrimination broke down in a more complex
crimination easier than does interspersed environment.
evocation. Although the roles of task complexity and
Demonstration of generalized imitation has similarity between model and subject remain
been easier than its explanation. Studies with somewhat obscure, the importance of rein-
normal children point to a number of vari- forcement is clearer (Bry & Nawas, 1972).
ables that may influence the performance of As Baer et al. (1967) and Peterson (1968)
nonreinforced imitative behavior. Burgess, have shown, reinforcement of some imitative
Burgess, and Esveldt (1970) suggested that behavior serves to maintain the general class
nonreinforced imitation occurs as a function of imitative behavior, and withdrawal of re-
of one (or a combination) of four variables: inforcement leads to the extinction of previ-
(a) the similarity between experimenter and ously reinforced, as well as nonreinforced,
subject takes on reinforcing value, (b) the imitative behavior.
subject imitates because of his interpretation The small number of well-executed studies
of the experimenter's instructions, (c) subtle using atypical children and the limited num-
cues from the experimenter reinforce behav- ber of subjects employed in these studies
ior, or (d) the subject discriminates (or fails make general statements about learning non-
to discriminate) situations in which he is verbal imitation difficult. Although subjects
reinforced from those in which he is not. do learn generalized imitation, at least within
Peterson and Whitehurst (1971) found, for topographically similar categories, the amount
example, that the presence or absence of the of time required and the degree of generaliza-
experimenter influenced the likelihood that tion to dissimilar categories vary. Parametric
their young subjects would imitate. studies examining variables that influence the
Subject characteristics such as intelligence speed of generalization and the limits of
may influence which of these factors is most transfer would have pragmatic value for the
important in a given instance. Bright chil- clinician.
dren, for example, may discriminate rein- Related to training of imitative behavior is
forced from nonreinforced behavior and then training to follow commands. While most imi-
respond on the basis of what they believe the tative training consists of a sequence in which
experimenter expects. In the case of the non- the experimenter says to the subject, Do
verbal child, many of whom are below aver- this, and then models a nonverbal behavior,
age intellectually and who lack the language in teaching obedience to commands the sub-
for complicated conceptualizations, a parsi- ject is told, Clap your hands, Stand up, and
monious explanation may lie in their failure so forth. The subject must make a vocal-
to discriminate reinforced from nonreinforced motor association, as compared to the visual-
behavior. The task of discrimination may motor association involved in nonverbal
actually be too complex for some of these imitation. While developmentally normal
children. children acquire the ability to imitate before
It might be instructive to create a situation they can follow commands, it is not clear
in which added stimulus cues would make it what implications this holds for teaching Ian-
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 569

guage to the nonverbal child, since the devel- cedures to enhance generalization of audi-
opmental model has not been shown to be tory—motor training is clear. It seems that on
superior for the training of these children. an intuitive basis visual-motor generalization
The value of teaching the following of com- may be simpler than auditory-motor because
mands, as compared to nonverbal imitation, in the first case, once the subject has learned
has not been fully explored nor has it been to relate his body to that of the model, he
shown in what sequence to place these two can see what he has to do, while in the
behaviors. Clinically, various approaches have second case he must understand all of the
been taken to the problem. Thus, Hewett words involved and make discriminations
(1965) mixed both imitation and command about the sequence of words before he re-
training during the same phase, Blake and sponds. Thus, while a few simple commands
Moss (1967) appear to have relied exclu- such as Look at me or Sit down may be
sively upon command training as their pre- quickly taught as specific instances, the gen-
cursor to verbal imitation, and the bulk of the eral training of commands may follow more
studies, as cited above, used imitation train- readily at a later stage of language training
ing but also included a few simple commands, than nonverbal imitation.
such as, "Look at me" or "Sit down." While
it is not clear which approach is most effi- VERBAL IMITATION
cient, the data suggest that generalized obe- Most researchers adopt some variation of
dience to commands follows later than gen- the four steps established by Lovaas et al.
eralized nonverbal imitation. This conclusion (1966) in training verbal imitation: (a) Re-
is based upon a few, but very consistent, ward all vocalization, (b) reward vocalization
studies. occurring within 6 seconds of the model's
Studies of following commands have gen- vocalization, (c) reward vocalization within
erally not been oriented toward eventual 6 seconds of the model's vocalization that
development of verbal behavior but have approximates the vocalization of the model,
focused instead upon the kind of nonverbal and (d) introduce a new sound randomly
obedience that is adaptive within an in- interspersed with the sound from Step c.
stitutional setting. Whitman, Zakaras, and Beyond this point the process is one of in-
Chardos (1971) trained two retarded subjects creasing the number of discriminated sounds.
to obey instructions and to generalize this Since the initial research of the mid-1960s
obedience to new instructions—although gen- (Hewett, 1965; Lovaas et al., 1966; Salzinger,
eralization was less likely to occur with more Feldman, Cowan, & Salzinger, 1965; Wolf,
complex commands or when the instructions Risley, & Mees, 1964; Risley & Wolf,
contained unfamiliar words. Two other impor- Note 1), there have been a variety of case
tant studies show the serious limits encoun- reports describing operant procedures for
tered during the generalization of compliance teaching speech to mute or echolalic children
to commands by nonverbal subjects. Striefel (e.g., Blake & Moss, 1967; Chapel, 1970;
and Wetherby (1973) found that although Colligan & Bellamy, 1968; Cook & Adams,
their subject learned to obey commands, this 1966; Goldstein & Lanyon, 1971; Guess,
behavior failed to generalize to untrained Rutherford, & Twichell, 1969; Kerr et al.,
probes. Their data suggest that the number 1965; Marshall & Hegrenes, 1970, 1972;
of words in the command, the position of the Picaizen, Berger, Baronofsky, Nichols, &
verb, and previous training with the noun Karen, 1969; Risley & Wolf, 1967; Schell,
and verb all influenced the subject's ability Stark, & Giddan, 1967; Sloane et al., 1968;
to respond to a new command. In a later Stark et al., 1968; Sulzbacher & Costello,
study, Striefel, Bryan, and Aikens (1974) 1970). These procedures have been applied in
transferred control of behavior from non- group as well as individual settings (e.g.,
verbal imitation to verbal instruction. Like Borus et al., 1973; Koegel & Rincover, 1974).
Striefel and Wetherby they reported a lack Reports on this work vary widely in sophisti-
of generalization to responses that had not cation and innovation. Nevertheless, repli-
been specifically trained. The need for pro- cating basic procedures in a variety of labora-
570 SANDRA L. HARRIS

tories increases the confidence one can place learning a verbal behavior but that once the
in the techniques. Since these procedures are behavior was established, social reinforcement
now well established, new case studies should was sufficient to maintain the response.
involve exploration of novel procedures and The forms of punishment used to suppress
should incorporate appropriate controls. The unwanted behaviors during language training,
need for continual evaluation is highlighted are quite variable and include isolation and
by the suggestion in the preceding section darkness (Hewett, 1965), time-out (e.g.,
that the value of nonverbal imitation training Borus et al., 1973; McReynolds, 1969), re-
may be largely "superstitious" in relation to sponse cost (McReynolds & Huston, 1971),
training verbal imitation. increased task complexity (Sailor, Guess,
As Jacobson, Bernal, and Lopez (1973) Rutherford, & Baer, 1968), and shouts and
have shown, this insistence upon experimen- slaps (Koegel & Rincover, 1974; Lovaas et
tal control is not an academic exercise. These al., 1966). The report by Steeves et al.
authors trained a retarded youth who had no (1970) that under certain conditions a self-
intelligible speech to make a nonverbal two- imposed time-out may be reinforcing, illus-
choice discrimination. Their subject learned trates once again the importance of careful
the discrimination, but of greater interest in observation of" each subject as well as defini-
the present context, he also started speaking tion of reinforcement and punishment on the
intelligible words. The skill was acquired basis of behavior change. While the majority
without direct verbal training. This finding of studies ignored errors and used punish-
suggests how essential it is that studies be ment only to reduce disruptive behavior,
designed to allow precise identification of McReynolds and Huston (1971-) punished
those factors that are in fact responsible for imitative errors by response cost. Their find-
the establishment of speech behavior. ing of decreased performance with response
A wide range of reinforcing agents have cost, as compared to reinforcement only,
been used to establish verbal behavior in suggests that their subjects may have had
nonverbal subjects. Although the most com- difficulty forming a discrimination between
mon of these is food coupled with praise, punishment for speech errors and speech
others have included colored lights (Blake & per se.
Moss, 1967; Fineman, 1968a, 1968b), tokens Generally in teaching verbal behavior, the
(e.g., Borus et al., 1967; Guess et al., 1969; model provides help for the subject in the
McReynolds & Huston, 1971; Steeves, form of prompting, fading, and other cues to
Martin, & Pear, 1970), music (Buddenhagen, responding. The form and extent of these cues
1971), physical contact (e.g., Kerr et al., has been variable. In a study designed to
1965; Lovaas et al., 1966), games (Hewett, develop vocalization in the presence of an-
1965; Sloane et al., 1968), and play with a other child, Hingtgen and Trost (1966) in-
tape recorder (Buddenhagen, 1971). Follow- creased the rate of vocalization between their
ing initial training on a continuous reinforce- subjects without any modeling or instructions
ment schedule, reinforcement is typically by the experimenter, but simply by shaping
shifted to an intermittent schedule although successive approximations of the desired re-
praise may remain on the continuous rein- sponse. However, the level of vocalization
forcement schedule (e.g., Drash, Caldwell, & that emerged was not as complex as that
Leibowitz, 1970; Salzinger et al., 1965; achieved in studies using more direct proce-
Schroeder & Baer, 1972; Steeves et al., 1970). dures. Sounds that permit manual prompting
This shift is critical because maintenance in and those that offer visual cues are often
the natural environment is almost always on selected for early training to maximize avail-
an intermittent basis. The importance of able cues (e.g., Lovaas et al., 1966; Sloane
using different kinds of reinforcement in dif- et al., 1968). Those studies using fading and
ferent phases of training was suggested by prompting generally describe a procedure of
McReynolds (1970), who demonstrated that forwarding chaining and fading (e.g., Yellow,
primary reinforcers were more effective than Yell, Ye). However, McReynolds (1967,
social reinforcers during the initial phases of 1970) and MacCubrey (1971) used backward
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 571

chaining to teach complex sounds (e.g., m, Evans (1968) used many physical prompts
im, rim, cream). There are no data com- in articulation training.
paring the differential effectiveness of these While most published articles describe suc-
two strategies. More importantly, there are cessful attempts to establish verbal imitative
no data indicating the long-term desirability behavior, one occasionally comes upon an
of providing prompts in language training, observation of failure (e.g., Garcia et al.,
nor that it is appropriate to reinforce 1971; Marshall & Hegrenes, 1972). One has
prompted behavior. Lovaas, Koegel, Simmons, to wonder about the rate of other failures,
and Long (1973) and Lovaas, Schreibman, which are unpublished. While it is possible to
and Koegel (1974) indicated that experience raise methodological questions about these
with their subject population led to serious studies, in both cases cited the authors were
reconsiderations of the use of prompts and established researchers who reported success
prompt-fading procedures. They suggested with other subjects. Thus, it does not seem
that stimulus overselectivity in autistic chil- likely that they did not know how to imple-
dren slows the, acquisition of responses where ment the procedures. Possibly there are sub-
a shift in stimulus control is required. Re- jects in whom biological damage prohibits
gardless of the theoretical explanation, the the learning of speech (although they are
pragmatic question of the use of prompts and doubtless rare). More likely, we still have a
the rewarding of prompted behavior remains good deal to learn about creating the proper
of critical clinical importance. Research in conditions for teaching speech. Reports of
this area must examine both short-term ef- failure might be very instructive for this
fects of learning on individual trials and purpose.
long-term effects in terms of retention and
generalization. FUNCTIONAL LANGUAGE
Hewett (1965), commenting on the diffi- Following the establishment of verbal imi-
culty of teaching his subject to form the tation, the main thrust of language training
discrimination between the first two sounds is aimed at functional language. Generally
he learned to imitate, noted that the use of the subject is first taught noun labels and
two different reinforcers may have aided in then other forms of grammar. There are,
the discrimination. Data from Schroeder and however, no data to aid in deciding the proper
Baer (1972) suggest that verbal imitation temporal sequence for introducing various
training is facilitated when the subject is grammatical forms. The only guidelines thus
trained on a set of words concurrently, as op- far are convenience and developmental norms.
posed to a serial procedure, in which the words Neither of these criteria has been demon-
are trained one at a time and discrimination strated to bear any direct relationship to the
must then be established. These data are learning needs of the nonverbal youngster.
compatible with Sailor and Taman's (1972) It is this stage that transforms the previous
demonstration that language-deficient autistic training into a useful skill—imitative speech
children had difficulty forming a discrimina- per se has no communication value.
tion between two prepositions taught with the Many case studies have reported attempts
same stimulus object. to teach functional language to nonverbal
Bricker and Bricker (1972), in evaluating subjects. Some of these involved children who
imitation errors made by retarded subjects, were initially mute (e.g., Brawley, Harris,
found that these errors were not random but Allen, Fleming, & Peterson, 1969; Colligan
instead contained parts of the correct re- & Bellamy, 1968; Hewett, 1965; Stark et al.,
sponses. Finding that the most common errors 1968), while others had limited or echolalic
were labial-dental and lingua-dental place- speech (e.g., Goldstein & Lanyon, 1971;
ments, they attempted to train one group in Jensen & Womack, 1967; Matheny, 1968;
these deficiencies, while a second group was McClure, 1968; Ney, 1967; Ney, Palvesky,
trained on each sound directly. Contrary to & Markely, 1971; Shaw, 1969; Sulzbacher &
expectation, there were no posttest differ- Costello, 1970; Tramontana & Shivers, 1971;
ences between the two groups. Nelson and Tramontana & Stimbert, 1970; Weiss &
 572 SANDRA L. HARRIS

Born, 1967; Wetzel, Baker, Roney, & Martin, two subjects received the opposite training.
1966; Wolf, Risky, & Mees, 1964). Although The words used were appropriate to the ending
all researchers reported some success in estab- being trained (e.g., hat/hats versus bus/
lishing functional speech, the degree of ex- busses). Generalization to the other modality
perimental control was highly variable; conse- (receptive or expressive) using the same
quently, it is difficult to identify the active plural ending (/-s/ or /-es/) was tested by
treatment ingredients. Of particular interest, nonreinforced probes. Thus, a subject trained
there is often little information about the on expressive use of /-s/ endings was periodi-
degree of support for verbal behavior available cally probed for his receptive use of /-s/
from the environment. endings. Consistent with previous research,
In addition to case reports, a series of em- Guess and Baer (1973a) found that teaching
pirical studies during the past few years have subjects to respond to novel instances within
explored the teaching of various grammatical the same response class was relatively easy
forms and the procedures for establishing gen- but that generalization across response classes
eralization of language. In one of the first of remained more difficult to establish. Interest-
these studies, Guess, Sailor, Rutherford, and ingly, however, one of the four subjects did
Baer (1968), using imitation and differential show this generalization without any addi-
reinforcement, taught expressive plural nouns tional training. The remaining three subjects
to a subject who had never used plural forms. were trained in the opposite response classes
The following year, Guess (1969) used simi- by reinforcement of probe responses. In
lar procedures to establish receptive auditory their discussion, Guess and Baer (1973a)
plurals in two retarded subjects. He demon- emphasized,
strated once again the relative ease of build-
ing the response class of plurals but found it is apparent that generalization between two mo-
that in spite of their receptive skill, the sub- dalities can occur, but by no means is an "automatic"
phenomenon, even in conditions . . . strongly em-
jects were unable to generalize the knowledge phasizing the functional similarity of both modalities,
to an expressive plural response class until (p, 328)
specifically trained in that class. It appears
that for some subjects expressive and recep- A more detailed description and a discussion
tive language may be functionally indepen- of the theoretical importance of these studies
dent. Sailor (1971) explored the extent to for language acquisition may be found in a
which differential reinforcement influenced a recent chapter by Guess and Baer (1973b).
subject's acquisition of various plural endings. Other studies that have built expressive or
The /-s/ and /-z/ endings were both acquired receptive response classes into the language
with similar ease, supporting the notion that of subjects of limited verbal ability include
differential reinforcement rather than a phys- the training of receptive prepositions (Frisch
iological factor is the crucial element in this & Schumaker, 1974), singular and plural de-
learning. Guess and Baer (1973a) attempted clarative sentences (Garcia, Guess, & Byrnes,
to remedy some of the faults of earlier re- 1973), verb tenses (Lutzker & Sherman,
search on the plural response class with a 1974; Schumaker & Sherman, 1970), adjec-
study in which generative plural formation tival inflections (Baer & Guess, 1971), com-
rules were taught concurrently for both ex- pound sentences (Stevens-Long & Rasmussen,
pressive and receptive language. Their study 1974), complex sentences (Odom, Liebert, &
merits additional description. Four severely Fernandez, 1969; Wheeler & Sulzer, 1970),
retarded subjects able to articulate the /-s/ and interrogative sentences (Twardosz & Baer,
and /-es/ sounds necessary to form plurals, 1973). Many of these studies point toward
but displaying neither expressive nor recep- the difficulty of teaching retarded or autistic
tive plural usage on pretest, were trained for subjects to generalize from, expressive to
30 minutes a day using tokens and praise on receptive language or vice versa, an im-
an intermittent schedule. Two subjects were portant finding whose implications extend
taught to use the /-s/ ending receptively and beyond the training of any particular gram-
the /-es/ ending expressively, while the other matical form. While it has been theorized
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 573

that receptive language precedes expres- trated by the work of Koegel and Rincover
sive language (Chomsky, 1967; Lennenberg, (1974), who found that behaviors learned in
1962; Mykelbust, 1957) and even demon- a one-to-one setting did not systematically
strated with normal subjects (Fraser, Bellugi, transfer to a one-to-eight classroom. Since a
& Brown, 1963; Mann & Baer, 1971; Winitz number of studies (Borus et al., 1973; Gray
& Preisler, 1965), this developmental model & Ryan, 1973; MacCubrey, 1971; Martin
may not hold true for the nonverbal child. et al., 1968) have demonstrated the feasibil-
The risk inherent in attempting to generalize ity of group training, it appears desirable
from the normal speaker to the nonverbal to ensure that at least part of a nonspeaker's
child thus becomes evident. training be done within a group to facilitate
Although the subjects in the studies by generalization.
Guess (1969) and Guess and Baer (1973a) One outgrowth of the research on teaching
failed to show "spontaneous" generalization functional language has been the publica-
to probes in the untrained (expressive or tion of systematic training programs (e.g.,
receptive) modality, both studies suggest that Blindert, 1971; Bricker, 1972; Bricker &
training in one modality may facilitate learn- Bricker, 1970; Gray & Ryan, 1973; Hartung,
ing in the other. Thus, Guess (1969) ob- 1970; Kent, 1974; Kent et al., 1972; Lovaas,
served that the expressive acquisition of sin- 1971; Marshall & Hegrenes, 1970; McKenna-
gulars and plurals was more rapid than the Hartung & Hartung, 1973; McLean &
earlier receptive training. Likewise, Guess and McLean, 1974; Miller & Yoder, 1972; Rosen,
Baer (1973a) noted that reinforcement of Wesner, & Zisfein, 1972; Sailor, Guess, &
probe responses in the untrained modality Baer, 1973; Stark, Rosenbaum, Schwartz, &
lead to the fairly rapid acquisition of the Wilson, 1973; Stremel, 1972). These pro-
correct response in that modality. These data grams vary widely in the empirical support
are encouraging because they suggest that they provide for the assumptions upon which
while responses across modalities are not auto- they are based, the extent to which they have
matic for all subjects, many show a savings been tested on clinical populations, prerequi-
with previous training in another modality. site language for entrance to the program,
For some subjects it appears that the task guidance in selecting criteria for advancing
is to learn not the content of the response from step to step in the program, and the
but the discriminative stimuli setting the amount of detail provided. For example,
occasion for the response. For example, once Sailor et al. (1973) observed that if Bricker's
learning that plural responses are called for (1972) model were fully carried out, it
in both expressive and receptive conditions, would surely produce functional language, but
they have available the knowledge learned it would be so lengthy to execute that it
from one modality for use in the other. would rarely be applied.
Once language has been established within Two language training programs built upon
the training situation, the next step is the a developmental model are Stremel's (1972)
extension of functional speech to other people and Miller and Yoder's (1972), both of
in addition to the trainer and to other settings which cite the developmental data of Bloom
beyond the therapy room. Many of the best- (1970). The Gray and Ryan (1973) pro-
designed studies have failed to attempt this gram, not built exclusively upon develop-
level of generalization. One exception is the mental assumptions, is highly elaborate in
recent study by Garcia (1974), who first detail and logical in construction but has been
trained his subjects in a simple conversational tested primarily with dysphasic children of
sequence and then probed for generalization normal intelligence (Fygetakis & Gray, 1970;
to other persons and settings. His data point Gray & Fygetakis, 1968a, 1968b). The pro-
to the importance of using more than one gram of Kent et al. (1972) focuses upon the
person to train subjects from the onset of early stages of language training including
training. nonverbal precursors, while that of Gray and
The desirability of training subjects to re- Ryan (1973) focuses more upon higher levels
spond in a variety of settings is also illus- of functioning and assumes some initial com-
574 SANDRA L. HARRIS

petence. In general these programs all require has had at least a limited amount of lan-
continued testing with a diversity of subjects. guage prior to training, "once some speech
There should be an emphasis in all of them begins to emerge irTthe presence of an audi-
upon generalization to the natural environ- ence, a great deal more that has not been
ment. specifically taught will emerge as well" (p.
The transition from a formal training set- 239). Lovaas et al. (1973) reported that
ting, be it individual or group, to the natural their mute subjects were more behaviorally
environment has received almost no system- retarded than their echolalic subjects at in-
atic evaluation although this is the end goal take, and while the mute subjects made pro-
of language training. Most of our data in this portionally greater gains, the echolalic sub-
area come from case reports. Weiss and Born jects still appeared qualitatively superior at
(1967), in an early report, observed that the follow-up. Although these data do not negate
generalization of functional language shown the value of speech to a child who would
by their subject was disappointingly small in have been mute without training, they do
spite of encouragement by the ward staff. In suggest that the existence of language prior
general, studies that set out to train parents to training may remain an important prog-
as therapists have reported good results in nostic index.
achieving generalization to the home and
school (e.g., Brawley et al., 1969; Hewett, Experimental Design
1965; Kozloff, 1973). Although difficulties inherent in this area
The data pointing most dramatically to the demand that a sympathetic view be taken of
importance of the environment in maintaining the problems confronting the researcher, much
language behavior taught within the clinical of the research regarding teaching language
setting are those in the follow-up report of to the nonverbal subject is open to criticism.
Lovaas et al. (1973), which clearly show the First, a number of different populations have
reversibility of treatment effects when the been subsumed under the umbrella term non-
environment fails to make appropriate de- verbal. Some subjects are echolalic, others
mands and reward desirable behavior. have limited functional speech, some are mute
and have never talked, others are presently
DISCUSSION mute but have a history of some language.
Among the most important issues to The subjects likewise vary in IQ estimates—
emerge from this review are questions of from "untestable" to normal intellectual po-
prognosis, experimental design, and generali- tential. The diagnostic labels include autism,
zation to the natural environment. childhood schizophrenia, dysphasia, mental
retardation, and brain damage. For some, en-
Prognosis vironmental trauma may be identified; for
Data suggest that children who fail to others there are no obvious precursors of de-
develop speech by 5 years of age have a ficient language. These factors produce a
poorer prognosis than do those who have highly heterogeneous population under the
language by that time (DeMyer, Barton, De- rubric of nonverbal and thereby compound
Myer, Norton, Allen, & Steele, 1973; Eisen- the researcher's job by making it hard to find
berg, 1956; Rutter & Lockyer, 1967a, 1967b). matched groups of subjects.
What influence will the use of operant pro- One response to this heterogeneity of sub-
cedures for teaching speech have upon that ject population has been to design most re-
prognostic factor? There is agreement among search using single subjects. This strategy,
researchers that it is easier to teach language using both reversal and multiple-baseline
to a child who is echolalic or has had pre- designs to employ each subject as his own
vious language than it is to teach a child who control, answers some of the methodological
has never spoken. Wolf, Risley, and Mees problems presented by a mixed population.
(1964) discussed the greater ease with which Nevertheless, it raises another question—-
their echolalic subjects learned to speak. the extent to which data based on one or two
Evans (1971) noted that for the child who subjects can be safely generalized to a wider
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 575

group. The need for larger studies with ran- mental demands in the training of the non-
domly selected subjects (Bricker, 1972) or verbal child.
for replications of single^subject studies is In situations in which the reversal design is
obvious. If the n = 1 study is repeated sev- ethically questionable or aesthetically unat-
eral times, this permits the identification of tractive, the multiple-baseline design provides
those variables that may influence behavior valuable experimental control. Here again,
in different subjects. Hence, it is not the use parents or staff can be involved in the learn-
of single-subject studies that is objectionable, ing experience and sensitized to the impor-
but the use of only one or two subjects to tance of maintaining proper demands upon
draw conclusions that are applied to a hetero- the child. Generally, people are responsive to
geneous population. Even within many of the notion that a given behavior, having ex-
the operant studies using two or three sub- isted for n amount of time, can be allowed to
jects, the authors described one subject who go on for n,+ 2 weeks (say) until other
failed to follow the pattern of the others. behaviors have been trained. Obviously the
These variations make it clear that we have span of time is a critical variable. Just as
not identified all of the parameters involved the duration of a reversal must be as brief as
in language training; they make broader possible, so too must be the amount of time
sampling and replication imperative. that treatment is withheld in a multiple base-
One fact that emerges from the examina- line. Few people would object to withholding
tion of the case studies reviewed here is that some aspect of training for a few weeks,
in many instances the authors excused their while many would be appropriately distressed
research on clinical grounds for failing to if training were withheld for months simply
provide even minimal controls, to collect to satisfy experimental demands. The nature
systematic data, or to determine the reliabil- of the behavior must likewise be considered.
ity of their observations. The use of rever- The withholding of treatment for a life-threat-
sals, even for a brief time, or of a multiple- ening behavior has different implications than
baseline design would strengthen many a not training the preposition on, while the
clinical report beyond the state of being sim- preposition in is trained.
ply an anecdotal description.
Generalization
Many psychologists have understandable
reservations about using the reversal design in One issue emerged again and again
a clinical setting. Nevertheless, to the best of throughout this review: the problem of teach-
the present author's knowledge there are no ing subjects to generalize their responses to
reports of studies in which a return to base- novel stimuli within the same response class
line, once instituted, could not be readily and to respond in a variety of settings and to
undone, with a rapid return to treatment lev- a variety of people. The work of Lovaas,
els of responding. To the contrary, the more Schreibman, Koegel, and Rehm (1971) re-
typical problem is to show the decrement in garding the responses of autistic, retarded,
performance that ideally accompanies the re- and normal children to multiple-stimulus in-
versal. It is often important to execute the put may be relevant to this problem. In
reversal early enough in training to ensure their study multiple stimuli (auditory, visual,
that responding has hot become so generalized and tactile) were presented simultaneously as
that extinction is difficult. In addition, an signals that bar pressing would be reinforced.
experimental necessity may be turned into a After responding was established, the subjects
were tested by the presentation of the various
therapeutic asset by using the reversal design
components of the multiple-stimulus complex.
to show parents or other caretakers precisely Lovaas et al. found that autistic subjects
how important the training procedures are for responded to only one of the stimuli, retarded
the maintenance of behavior. An effective subjects to two, and normal subjects to all
reversal can illustrate for the lay person in a three. What stands out, in the context of the
concrete and dramatic fashion the impor- present discussion, is that neither autistic nor
tance of maintaining appropriate environ- retarded subjects were able to use all of the
 576 SANDRA L. HARRIS

cues available in the environment to mark zation into the home. In H. Work (Chair),
the time for correct responding. A similar Experimental studies in childhood schizophrenia.
phenomenon may be noted in the language Symposium presented at the meeting of the
American Psychological Association, Los Angeles,
behavior of both groups. Having learned lan- September 1964.
guage under highly controlled conditions, they
often failed to generalize to conditions that
contained some, but not all, of the cues present REFERENCES
during initial training. Since the autistic chil- Baer, D. M., & Guess, D. Receptive training of
dren in the Lovaas et al. research were also adjectival inflection in mental retardates. Journal
the most intellectually deficient, it may be of Applied Behavior Analysis, 1971, 4, 129-139.
Baer, D. M., Peterson, R., & Sherman, J. The
that the important dimension was IQ rather development of imitation by reinforcing behavioral
than the phenomenon of stimulus overselectiv- similarity to the model. Journal of Experimental
ity that is linked with autism. This 'more Analysis of Behavior, 1967, 10, 405-416.
parsimonious explanation would cut across Bandura, A. Social-learning theory of identifactory
processes. In D. A. Goslin & D. D. Glass (Eds.),
diagnostic groups and account for generaliza- Handbook of socialization theory and research.
tion difficulties in heterogeneous subjects. Chicago: Rand McNally, 19.68.
If intellectually retarded nonverbal people Blake, P., & Moss, T. The development of socializa-
have particular difficulty forming necessary tion skills in an electively mute child. Behaviour
discriminations and identifying the cues that Research and Therapy, 1967, 5, 349-356.
Blindert, H. D. The science of behavior, behavior
set the occasion for different responses, one modification, and verbal behavior. IRAL: Inter-
of the important factors in facilitating gen- national Review of Applied Linguistics, 1971, 9,
eralization is repeated training in a wide 53-62.
number of settings. This means that language Bloom, L. Language development: Form and junc-
tion in emerging grammars. Cambridge, Mass.:
training must focus upon the natural environ- MIT Press, 1970.
ment rather than the classroom. Borus, J. F., Greenfield, S., Siegel, B., & Daniels, G.
This review has shown that it is relatively Establishing imitative speech employing operant
easy to establish generalized responding with- techniques in a group setting. Journal of Speech
in a given response class, be it gross motor and Hearing Disorders, 1973, 38, 533-541.
Brawley, E. R., Harris, F. R., Allen, K. E., Flem-
imitation, verbal imitation, or a given gram- ing, R. S., & Peterson, R. F. Behavior modifica-
matical construct. However, the transition tion of an autistic child. Behavioral Science, 1969,
from this stage to performance under wide- 14, 87-97.
ranging conditions has received far less at- Bricker, W. A. A systematic approach to language
training. In R. L. Schiefelbusch (Ed.), Language
tention than it deserves. This doubtless re- of the mentally retarded. Baltimore, Md.: Uni-
flects, at least in part, the greater difficulty versity Park Press, 1972.
that gathering data within the natural en- Bricker, W. A., & Bricker, D. D. A program of
vironment presents, as compared to the lab- language training for the severely handicapped
child. Exceptional Children, 1970, 37, 101-111.
oratory. Nevertheless, the technology for Bricker, W. A., & Bricker, D. D. Assessment and
doing such research exists. It appears that modification of verbal imitation with low-func-
this is the critical direction our attention t'.oning retarded children. Journal of Speech and
must now take. The training of parents as Hearing Research, 1972, 15, 690-698.
therapists is doubtless especially important in Brooks, B. D., Morrow, J. E., & Gray, W. F. Re-
duction of autistic gaze aversion by reinforcement
this respect (e.g., Goldstein & Lanyon, 1971; of visual attention responses. Journal of Special
Hewett, 1965; Kozloff, 1973). Lovaas et al. Education, 1968, 2, 307-309.
(1973), in their follow-up study, made it clear Bry, P. M., & Nawas, M. M. Is reinforcement
that we must facilitate the transfer of the necessary for the development of a generalized
imitation operant in severely and profoundly
child's skills to the natural environment and retarded children? American Journal of Mental
ensure that there is sufficient reinforcement Deficiency, 1972, 76, 658-667.
to sustain these behaviors beyond the therapy Buddenhagen, R. Establishing vocalizations in mute
room. mongoloid children. Champaign, 111.: Research
Press, 1971.
REFERENCE NOTE Burgess, R. L., Burgess, J. M., & Esveldt, K. C.
1. Risley, T. R., & Wolf, M. M. Experimental An analysis of generalized imitation. Journal of
manipulation of autistic behaviors and generali- Applied Behavior Analysis, 1970, 3, 39-46.
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 577

Chapel, J. L. Behaviour modification techniques cally determined boundaries. Journal of Applied

with children and adolescents. Canadian Psychi- Behavior Analysis, 1971, 4, 101-112.
atric Association Journal, 1970, 15, 315-318. Garcia, E., Guess, D., & Byrnes, J. Development of
Chomsky, N. The general properties of language. syntax in a retarded girl using procedures of imi-
In C. Millikan & F. Darley (Eds.), Brain mech- tation, reinforcement, and modeling. Journal of
anisms underlying speech and language. New Applied Behavior Analysis, 1973, 6, 299-310.
York: Grune & Stratton, 1967. Gardner, J. E., Pearson, D. T., Bercovici, A. N., &
Churchill, D. W. Psychotic children and behavior Bricker, D. E. Measurement, evaluation, and
modification. American Journal of Psychiatry, modification of selected social interactions between
1969, 125, 1585-1S90. a schizophrenic child, his parents, and his thera-
Colligan, R. C., & Bellamy, C. M. Effects of a two- pist. Journal of Consulting and Clinical Psychol-
year treatment program for a young autistic ogy, 1968, 32, 537-542.
child. Psychotherapy: Theory, Research and Goldstein, S. B., & Lanyon, R. I. Parent-clinicians
Practice, 1968, 5, 214-219. in the language training of an autistic child.
Cook, C., & Adams, H. E. Modification of verbal Journal of Speech and Hearing Disorders, -1971,
behavior in speech deficient children. Behaviour 36, 552-560.
Research and Therapy, 1966, 4, 265-271. Gray, B. B., & Fygetakis, L. The development of
DeMyer, M. K., Barton, S., DeMyer, W. E., Nor- language as a function of programmed condition-
ton, J. A., Allen, J., & Steele, R. Prognosis in ing. Behaviour Research and Therapy, 1968, 6,
autism: A follow-up study. Journal of Autism 455-460. (a)
and Childhood Schizophrenia, 1973, 3, 199-246. Gray, B. B., & Fygetakis, L. Mediated language
Drash, P. W., Caldwell, L. R., & Leibowitz, J. M. acquisition for dysphasic children. Behaviour Re-
Correct and incorrect response rates as basic de- search and Therapy, 1968, 6, 263-280. (b)
pendent variables in the operant conditioning of Gray, B. B., & Ryan, B. A language training pro-
speech in non-verbal subjects. Psychological As- gram for the non-language child. Champaign, 111.:
pects of Disability, 1970, 17, 16-23. Research Press, 1973.
Eisenberg, L. The autistic child in adolescence. Guess, D. A functional analysis of receptive lan-
American Journal of Psychiatry, 1956, 112, 607- guage and productive speech: Acquisition of the
613. plural morpheme. Journal of Applied Behavior
Evans, I. M. Theoretical and experimental aspects Analysis, 1969, 2, 55-64.
of the behaviour modification approach to autistic Guess, D., & Baer, D. M. An analysis of individual
children. In M. Rutter (Ed.), Infantile autism: differences in generalization between receptive and
Concepts, characteristics and treatment. Baltimore, productive language in retarded children. Journal
Md.: Williams & Wilkins, 1971. of Applied Behavior Analysis, 1973, 6, 311-329.
Fineman, K. R. Shaping and increasing verbaliza- (a)
tions in an autistic child in response to visual- Guess, D., & Baer, D. M. Some experimental analy-
color stimulation. Perceptual and Motor Skills, ses of linguistic development in institutionalized
1968, 27, 1071-1074. (a) retarded children. In B. B. Lahey (Ed.), The
Fineman, K. R. Visual-color reinforcement in estab- modification of language behavior. Springfield, 111.:
lishment of speech by an autistic child. Perceptual Charles C Thomas, 1973. (b)
and Motor Skills, 1968, 26, 761-762. (b) Guess, D., Rutherford, G., & Twichell, A. Speech
Fineman, K. R., & Ferjo, J. Establishing and in- acquisition in a mute, visually impaired adoles-
creasing verbalizations in a deaf schizophrenic cent. New Outlook for the Blind, 1969, 63, 8-13.
child through the use of contingent visual-color Guess, D., Sailor, W., Rutherford, G., & Baer, D. M.
reinforcement. Perceptual and Motor Skills, 1969, An experimental analysis of linguistic develop-
29, 647-652. ment: The productive use of the plural morpheme.
Fraser, C., Bellugi, U., & Brown, R. Control of Journal of Applied Behavior Analysis, 1968, 1,
grammar in imitation, comprehension, and pro- 297-306.
duction. Journal of Verbal Learning and Verbal Harris, S. L. On the importance of being contingent
Behavior, 1963, 2, 121-135. —A reply to Ney. Journal of Autism and Child-
Frisch, S. A., & Schumaker, J. B. Training general- hood Schizophrenia, 1974, 4, 94-97.
ized receptive prepositions in retarded children. Hartung, J. R. A review of procedures to increase
Journal of Applied Behavior Analysis, 1974, 7, verbal imitation skills and increase functional
611-621. speech in autistic children. Journal of Speech and
Fygetakis, L., & Gray, B. Programmed conditioning Hearing Disorders, 1970, 35, 203-217.
of linguistic competence. Behaviour Research and Hewett, F. M. Teaching speech to an autistic child
Therapy, 1970, 8, 153-163. through operant conditioning. American Journal
Garcia, E. The training and generalization of a con- of Orthopsychiatry, 1965, 35, 927-936.
versational speech form in nonverbal retardates. Hingtgen, J. N., & Churchill, D. W. Identification
Journal of Applied Behavior Analysis, 1974, 7, of perceptual limitations in mute autistic children:
137-149. Identification by the use of behavior modifica-
Garcia, E., Baer, D. M., & Firestone, I. The devel- tion. Archives of General Psychiatry, 1969, 21,
opment of generalized imitation within topographi- 68-71.
578 SANDRA L. HARRIS

Hingtgen, J. N., & Churchill, D. W. Differential development of complex behavior in schizophrenic

effects of behavior modification in four mute au- children. Behaviour Research and Therapy, 1967,
tistic boys. In D. W. Churchill, G. D. Alpern, & 5, 171-182.
M. K. DeMyer (Eds.), Infantile autism. Spring- Lovaas, O. L, Koegel, R., Simmons, J. Q., '& Long,
field, 111.: Charles C Thomas, 1970. J. S. Some generalization and follow-up measures
Hingtgen, J. N., Coulter, S. K., & Churchill, D. W. on autistic children in behavior therapy. Journal
- Intensive reinforcement of imitative behavior in of Applied Behavior Analysis, 1973, 6, 131-166.
mute, autistic children. Archives of 'General Psy- Lovaas, O. L, Schreibman, L., & Koegel, R. L. A
chiatry, 1967, 17, 36-43. behavior modification approach to the treatment
Hingtgen, J. N., & Trost, F. C. Shaping cooperative of autistic children. Journal of Autism and Child-
responses in early childhood schizophrenics: II. hood Schizophrenia, 1974, 4, 111-129.
Reinforcement of mutual physical contact and Lovaas, O. L, Schreibman, L., Koegel, R., & Rehm,
vocal responses. In R. Ulrich, X. Stachnick, & R. Selective responding by autistic children to
J. Mabry (Eds.), Control of human behavior multiple sensory input. Journal of Abnormal Psy-
(Vol. 1). Glenview, 111.: Scott, Foresman, 1966. chology, 1971, 77, 211-222.
Isaacs, W., Thomas, J., & Goldiamond, I. Applica- Lutzker, J. R., & Sherman, J. A. Producing genera-
tion of operant conditioning to reinstate verbal tive sentence usage by imitation and reinforcement.
behavior in mute psychotics. Journal of Abnormal Journal of Applied Behavior Analysis, 1974, 7,
Psychology, 1965, 70, 155-164. 447-460.
Jacobson, L. I., Bernal, G., & Lopez, G. W. Effects MacCubrey, J. Verbal operant conditioning with
of behavioral training on the functioning of a young institutionalized Down's syndrome children.
profoundly retarded microcephalic teenager with American J-ournal of Mental Deficiency, 1971, 75,
cerebral palsy and without language or verbal 696-701.
comprehension. Behaviour Research and Therapy, Mann, R. A., & Baer, D. M. The effects of receptive
1973, 11, 143-145. language training on articulation. Journal of
Jensen, G. D., & Womack, M. Operant condition- Applied Behavior Analysis, 1971, 4, 291-298.
ing techniques in the treatment of an autistic Marshall, N. R., & Hegrenes, J. R. Programmed
child. American Journal of Orthopsychiatry, 1967, communication therapy for autistic mentally re-
37, 30-34. tarded children. Journal of Speech and- Hearing
Kent, L. R. Language acquisition program for the Disorders, 1970, 35, 70-83.
severely retarded. Champaign, 111.: Research Press, Marshall, N. R., & Hegrenes, J. The use of written
1974. language as a communication system for an au-
Kent, L. R., Klein, D., Falk, A., & Guenther, H. A tistic child. Journal of Speech and Hearing Dis-
language acquisition program for the retarded. In orders, 1972, 37, 258-261.
J. E. McLean, D. E. Yoder, & R. L. Schiefel- Martin, G. L., England, G., Kaprowy, E., Kilgour,
bush (Eds.), Language intervention with the K., & Pilek, V. Operant conditioning of kinder-
retarded. Baltimore, Md.: University Park Press, garten class behavior in autistic children. Be-
1972. haviour Research and Therapy, 1968, 6, 281-294.
Kerr, N., Meyerson, L., & Michael, J. A procedure Matheny, A. B. Pathological echoic responses in a
for shaping vocalizations in a mute child. In L. child: Effect of environmental mand and tact
P. Ullmann & L. Krasner (Eds.), Case studies in control. Journal of Experimental Child Psychol-
behavior modification. New York: Holt, Rine- ogy, 1968, 6, 624-631.
hart & Winston, 1965. McClure, R. Reinforcement of verbal social be-
Koegel, R. L., & Rincover, A. Treatment of psy- havior in moderately retarded children. Psycho-
chotic children in a classroom environment: I. logical Reports, 1968, 23, 371-376.
Learning in a large group. Journal of Applied McConnell, O. L. Control of eye contact in an au-
Behavior Analysis, 1974, 7, 45-60. tistic child. Journal of Child Psychology and
Kozloff, M. Reaching the autistic child. A parent Psychiatry and Allied Disciplines, 1967, 8, 249-255.
training program. Champaign, 111.: Research Press, McKenna-Hartung, S., & Hartung, J. R. Establish-
1973. ing verbal imitation skills and functional speech
Lennenberg, E. H. Understanding language without in autistic children. In B. B. Lahey (Ed.), The
ability to speak: A case report. Journal of Ab- modification of language behavior. Springfield, 111.:
normal and Social Psychology, 1962, 65, 419-425. Charles C Thomas, 1973.
Lovaas, 0. I. Considerations in the development of McLean, L. P., & McLean, J. E. A language train-
a behavioral treatment program for psychotic ing program for nonverbal autistic children. Jour-
children. In D. W. Churchill, G. D. Alpern, & nal of Speech and Hearing Disorders, 1974, 39,
M. K. DeMyer (Eds.), Infantile autism. Spring- 186-193.
field, 111.: Charles C Thomas, 1971. McReynolds, L. V. Application of timeout from
Lovaas, O. I., Berberich, J. P., Perloff, B. F., & positive reinforcement for increasing the efficiency
Schaeffer, B. Acquisition of imitative speech by of speech training. Journal of Applied Behavior
schizophrenic children. Science, 1966, 151, 705-707. Analysis, 1969, 2, 199-205.
Lovaas, O. I., Freitas, L., Nelson, K., & Whalen, C. McReynolds, L. V. Reinforcement procedures for
The establishment of imitation and its use for the establishing and maintaining echoic speech by a
 TEACHING LANGUAGE TO NONVERBAL CHILDREN 579

nonverbal child. Applications of a functional ap- Rutter, M., & Lockyer, L. A five to fifteen year
proach to speech and hearing. In F. Giradeau & follow-up study of infantile psychosis: II. Social
J. Spradlin (Eds.), ASHA Monographs, 1970, 14, and behavioural outcome. British Journal of Psy-
60-66. chiatry, 1967, 113, 1183-1199. (b)
McReynolds, L. V., & Huston, K. Token loss in Sailor, W. Reinforcement and generalization of
speech imitation training. Journal of Speech and productive plural allomorphs in two retarded chil-
Hearing Disorders, 1971, 36, 486-49S. dren. Journal of Applied Behavior Analysis, 1971,
Metz, J. R. Conditioning generalized imitation in 4, 305-310.
autistic children. Journal of Experimental Child Sailor, W., Guess, D., & Baer, D. M. Functional
Psychology, 1965, 2, 389-399. language for verbally deficient children. Mental
Miller, J. F., & Yoder, D. E. A language acquisition Retardation, 1973, 11, 27-35.
program for the retarded. In J. E. McLean, D. Sailor, W., Guess, D., Rutherford, G., & Baer, D. M.
E. Yoder, & R. L. Schiefelbusch (Eds.), Language Control of tantrum behavior by operant tech-
intervention with the retarded. Baltimore, Md.: niques during experimental verbal training. Jour-
University Park Press, 1972. nal of Applied Behavior Analysis, 1968, 1, 237-
Mykelbust, H. R. Auditory disorders in children. 243.
New York: Grune & Stratton, 19S7. Sailor, W., & Taman, T. Stimulus factors in the
Nelson, R. O., & Evans, I. M. The combination of training of prepositional usage in three autistic
learning principles and speech therapy techniques children. Journal of Applied Behavior Analysis,
in the treatment of non-communicating children. 1972, 5, 183-190.
Journal of Child Psychology and Psychiatry and
Allied Disciplines, 1968, 9, 111-124. Salzinger, K., Feldman, R. S., Cowan, J. E., &
Ney, P. Operant conditioning of schizophrenic chil- Salzinger, S. Operant conditioning of verbal be-
dren. Canadian Psychiatric Association Journal, havior of two young speech-deficient boys. In L.
1967, 12, 9-15. Krasner & L. P. Ullmann (Eds.), Research in
Ney, P. G. Effects of contingent and non-contingent behavior modification. New York: Holt, Rinehart
reinforcement on the behavior of an -autistic child. & Winston, 1965.
Journal of Autism and Childhood Schizophrenia, Schell, R. E., Stark, J., & Giddan, J. G. Develop-
1973, 3, 115-127. ment of language behavior in an autistic child.
Ney, P. G., Palvesky, A. E., & Markley, J. Relative Journal of Speech and Hearing Disorders, 1967,
effectiveness of operant conditioning and play 32, 51-64.
therapy in childhood schizophrenia. Journal of Schroeder, G., & Baer, D. M. Effects of concurrent
Autism and Childhood Schizophrenia, 1971, 1, 337- and serial training on generalized vocal imitation
349. in retarded children. Developmental Psychology,
Odom, R. D., Liebert, R. M., & Fernandez, L. Ef- 1972, 6, 293-301.
fects of symbolic modeling on the syntactical pro- Schumaker, J., & Sherman, J. A. Training genera-
ductions of retardates. Psychondmic Science, 1969, tive verb usage by imitation and reinforcement
17, 104-105. procedures. Journal of Applied Behavior Analysis,
Peterson, R. F. Some experiments on the organiza- 1970, 3, 273-287.
tion of a class of imitative behaviors. Journal of Shaw, W. H. Treatment of schizophrenic speech
Applied Behavior Analysis, 1968, 1, 225-235. disorder by operant conditioning in play therapy.
Peterson, R. F., & Whitehurst, G. J. A variable Canadian Psychiatric Association Journal, 1969,
influencing the performance of generalized imi- 14, 631-634.
tative behaviors. Journal of Applied Behavior Sherman, J. A. Reinstatement of verbal behavior in
Analysis, 1971, 4, 1-9. a psychotic by reinforcement methods. Journal of
Picaizen, G., Berger, A. A., Baronofsky, D., Nichols, Speech and Hearing Disorders, 1963, 28, 398-401.
A. C., & Karen, R. Applications of operant tech- Sherman, J. A. Use of reinforcement and imitation
niques to speech therapy with nonverbal children. to reinstate verbal behavior in mute psychotics.
Journal of Communication Disorders, 1969, 2, Journal of Abnormal Psychology, 1965, 70, 155-
203-211. 164.
"Risley, T. R. The effects and side effects of punish- Sloane, H. N., Johnston, M. K., & Harris, F. R.
ing the autistic behaviors of a deviant child. Remedial procedures for teaching verbal be-
Journal of Applied Behavior Analysis, 1968, 1, havior to speech deficient or defective young chil-
21-34. dren. In H. N. Sloane & B. D. MacAulay (Eds.),
Risley, T. R., & Wolf, M. M. Establishing func- Operant procedures in remedial speech and lan-
tional speech in echolalic children. Behaviour guage training. Boston: Houghton Mifflin, 1968.
Research and Therapy, 1967, 5, 73-88. Stark, J., Giddan, J. J., & Meisel, J. Increasing
Rosen, M., Wesner, C., & Zisfein, L. Your child can verbal behavior in an autistic child. Journal of
talk too. Elwyn, Pa.: Elwyn Institute, 1972. Speech and Hearing Disorders, 1968, 33, 42^8.
Rutter, M., & Lockyer, L. A five to fifteen year Stark, J., Rossnbaum, R. L., Schwartz, D., & Wilson,
follow-up study of infantile psychosis: I. Descrip- A. The nonverbal child: Some clinical guidelines.
tion of the sample. British Journal of Psychiatry, Journal of Speech and Hearing Disorders, 1973,
1967, 113, 1169-1182. (a) -38, 59-71.
580 SANDRA L. HARRIS

Sleeves, J. M., Martin, G. L., & Pear, J. J. Self- Twardosz, S., & Baer, D. M. Training two severely
imposed time-out by autistic children during an retarded adolescents to ask questions. Journal of
operant training program. Behavior Therapy, 1970, Applied Behavior Analysis, 1973, 6, 655-661.
1, 371-381. Weiss, H. H., & Born, B. Speech training or lan-
Stevens-Long, J., & Rasmussen, M. The acquisition guage acquisition ? A distinction when speech
of simple and compound sentence structure in an training is taught by operant conditioning pro-
autistic child. Journal of Applied Behavior Analy- cedures. American Journal of Orthopsychiatry,
sis, 1974, 7, 473-479. 1967, 37, 49-55.
Stremel, K. Language training: A program for Wetzel, R. J., Baker, J., Roney, M., & Martin, M.
retarded children. Mental Retardation, 1972, 10, Outpatient treatment of autistic behavior. Be-
47-49. haviour Research and Therapy, 1966, 4, 169-177.
Striefel, S., Bryan, K. S., & Aikens, D. A. Transfer Wheeler, A. J., & Sulzer, B. Operant training and
of stimulus control from motor to verbal stimuli. generalization of a verbal response form in a
Journal of Applied Behavior Analysis, 1974, 7, speech-deficient child. Journal of Applied Be-
123-136. havior Analysis, 1970, 3, 139-147.
Striefel, S., & Wetherby, B. Instruction-following Whitman, T. L., Zakaras, M., & Chardos, S. Effects
behavior of a retarded child and its controlling of reinforcement and guidance procedures on in-
stimuli. Journal of Applied Behavior Analysis, struction-following behavior of severely retarded
1973, 6, 663-670. children. Journal of Applied Behavior Analysis,
Sulzbacher, S. I., & Costello, J. M. A behavioral 1971, 4, 283-290.
strategy for language training of a child with Winitz, H., & Preisler, L. Discrimination pretrain-
autistic behavior. Journal of Speech and Hearing ing and sound learning. Perceptual and Motor
Disorders, 1970, 35, 256-276. Skills, 1965, .20, 905-916.
Tramontana, J., & Shivers, O. Behavior modification Wolf, M. M., Risley, T. R., & Mees, H. Application
with an echolalic child: A case note. Psychological of operant conditioning procedures to the behavior
Reports, 1971, 29, 1034. problems of an autistic child. Behaviour Research
Tramontana, J., & Stimbert, V. E. Some techniques and Therapy, 1964, 1, 305-312.
of behavior modification with an autistic child.
Psychological Reports, 1970, 27, 498. (Received July 23, 1974)