Citation
Operant behavior in the human infant

Material Information

Title:
Operant behavior in the human infant differentiation and discrimination
Creator:
Deitz, Samuel Michael, 1945- ( Dissertant )
Newell, John M. ( Thesis advisor )
Ware, William B. ( Reviewer )
Jester, Robert E. ( Reviewer )
Pennypacker, H. S. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1971
Language:
English
Physical Description:
x, 79 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Child psychology ( jstor )
Discrimination learning ( jstor )
Infants ( jstor )
Learning ( jstor )
Legs ( jstor )
Mental stimulation ( jstor )
Sensory discrimination ( jstor )
Smiles ( jstor )
Social reinforcement ( jstor )
Visual stimuli ( jstor )
Dissertations, Academic -- Foundations of Education -- UF ( lcsh )
Foundations of Education thesis Ph. D ( lcsh )
Infants ( lcsh )
Operant behavior ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Four separate experimients were completed examining operant response differentiation and stimulus discrimination in human infants. The purpose of the studies was to determine whether or not infants could learn these tasks and under what circumstances were they most easily learned. Each experiment used one infant between the ages of six and fourteen weeks of age as its subject, the same basic apparatus, and visual stimuli as reinforcers. The response studied v/as left and right leg-movennents which v/ere defined as tense-extend motions of either leg which covered a distance of at least one inch. Left and right leg-movements were measured independently by the apparatus employed in these studies. This apparatus also controlled the presentation or removal of the visual reinforcers and the discriminative stimuli. Visual reinforcement consisted of either light flashes or movement of a reflecting mobile on a conjugate reinforcement schedule; the discriminative stimuli were both visual and audible. Baseline, experimental, and extinction phases occurred in all studies except Experiment I where final extinction data were unobtainable. The experimental phases differed in each study. Experiment I examined differentiated responding under a stimulus discrimination paradigm in one male infant using conjugate mobile reinforcement and audible discriminative stimuli. Experiment II was identical to Experiment I except visual discriminative stimuli were substituted for audible stimuli. Experiment III examined response differentiation in one female infant using light flashes as the reinforcement. A conditioning, extinction, reconditioning design was used in this experiment. Experiment IV was identical to Experiment III except that one male infant was used and the light flash reinforcement was increased in intensity. The results of these studies showed conditioning developed in half of the eight experimental conditions and was approached in two others. Appropriate discriminative responding developed under both conditions in Experiment I but developed under neither condition in Experiment II. In Experiment II, however, one condition approached appropriate responding. Differentiated responding developed under only one of the two conditions in both Experiments III and IV. It was found that the leg- movement preference demonstrated during baseline was an important determinant of appropriate responding in all experiments. Enhancement of that preference effected the only conditioning in the differentiation studies. In the discrimination studies reversal of that preference effected more appropriate responding, in Experiment 11, it was under this reversal contingency that conditioning was approached. The conclusions drawn from these studies included comments on potential social and medical benefits as well as statements concerned with future research on infant learning. It v/as mentioned that one important problem for researchers of infant learning is the discovery of reliable and powerful reinforcers. Visual stimuli seem to lack this effectiveness and the class of consumable reinforcers was suggested as the best possibility. An important potential benefit discussed was that contingency experiences such as those incorporated in these steadies may be of significant aid in building young infants' adaptive behavioral abilities. It was also suggested that investigation into left-right discrimination, in light of medical evidence currently available, may provide for early identification, prevention, and cure of behavioral deficits. Finally, it was concluded that, in addition to these potential benefits, investigations into infant learning can help build a complete theory of human development.
Thesis:
Thesis--University of Florida, 1971.
Bibliography:
Bibliography: leaves 74-78.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029485447 ( AlephBibNum )
AEG9250 ( NOTIS )
014373844 ( OCLC )

Downloads

This item has the following downloads:

PDF ( 3 MBs ) ( .pdf )

operantbehaviori00deitrich_Page_85.txt

operantbehaviori00deitrich_Page_14.txt

operantbehaviori00deitrich_Page_11.txt

operantbehaviori00deitrich_Page_63.txt

operantbehaviori00deitrich_Page_66.txt

operantbehaviori00deitrich_Page_84.txt

operantbehaviori00deitrich_Page_39.txt

operantbehaviori00deitrich_Page_74.txt

operantbehaviori00deitrich_Page_15.txt

operantbehaviori00deitrich_Page_86.txt

operantbehaviori00deitrich_Page_68.txt

operantbehaviori00deitrich_Page_45.txt

operantbehaviori00deitrich_Page_46.txt

operantbehaviori00deitrich_Page_47.txt

operantbehaviori00deitrich_Page_22.txt

operantbehaviori00deitrich_Page_29.txt

operantbehaviori00deitrich_Page_13.txt

operantbehaviori00deitrich_Page_54.txt

operantbehaviori00deitrich_Page_44.txt

operantbehaviori00deitrich_Page_79.txt

operantbehaviori00deitrich_Page_48.txt

operantbehaviori00deitrich_Page_91.txt

operantbehaviori00deitrich_Page_01.txt

operantbehaviori00deitrich_Page_19.txt

operantbehaviori00deitrich_Page_78.txt

operantbehaviori00deitrich_Page_81.txt

operantbehaviori00deitrich_Page_80.txt

operantbehaviori00deitrich_Page_18.txt

operantbehaviori00deitrich_Page_50.txt

operantbehaviori00deitrich_Page_52.txt

operantbehaviori00deitrich_Page_24.txt

operantbehaviori00deitrich_Page_28.txt

operantbehaviori00deitrich_Page_73.txt

operantbehaviori00deitrich_Page_25.txt

operantbehaviori00deitrich_Page_55.txt

operantbehaviori00deitrich_Page_64.txt

operantbehaviori00deitrich_Page_51.txt

operantbehaviori00deitrich_Page_34.txt

operantbehaviori00deitrich_Page_31.txt

operantbehaviori00deitrich_Page_26.txt

operantbehaviori00deitrich_Page_16.txt

operantbehaviori00deitrich_pdf.txt

operantbehaviori00deitrich_Page_82.txt

operantbehaviori00deitrich_Page_30.txt

operantbehaviori00deitrich_Page_42.txt

operantbehaviori00deitrich_Page_92.txt

operantbehaviori00deitrich_Page_49.txt

operantbehaviori00deitrich_Page_04.txt

operantbehaviori00deitrich_Page_69.txt

operantbehaviori00deitrich_Page_71.txt

operantbehaviori00deitrich_Page_32.txt

operantbehaviori00deitrich_Page_89.txt

operantbehaviori00deitrich_Page_09.txt

operantbehaviori00deitrich_Page_75.txt

operantbehaviori00deitrich_Page_20.txt

operantbehaviori00deitrich_Page_72.txt

operantbehaviori00deitrich_Page_88.txt

operantbehaviori00deitrich_Page_67.txt

operantbehaviori00deitrich_Page_03.txt

operantbehaviori00deitrich_Page_37.txt

operantbehaviori00deitrich_Page_12.txt

operantbehaviori00deitrich_Page_17.txt

operantbehaviori00deitrich_Page_27.txt

operantbehaviori00deitrich_Page_61.txt

operantbehaviori00deitrich_Page_59.txt

operantbehaviori00deitrich_Page_40.txt

operantbehaviori00deitrich_Page_06.txt

operantbehaviori00deitrich_Page_43.txt

operantbehaviori00deitrich_Page_21.txt

operantbehaviori00deitrich_Page_76.txt

operantbehaviori00deitrich_Page_02.txt

operantbehaviori00deitrich_Page_23.txt

operantbehaviori00deitrich_Page_05.txt

operantbehaviori00deitrich_Page_53.txt

operantbehaviori00deitrich_Page_57.txt

operantbehaviori00deitrich_Page_70.txt

operantbehaviori00deitrich_Page_10.txt

operantbehaviori00deitrich_Page_83.txt

operantbehaviori00deitrich_Page_07.txt

operantbehaviori00deitrich_Page_41.txt

operantbehaviori00deitrich_Page_77.txt

operantbehaviori00deitrich_Page_60.txt

operantbehaviori00deitrich_Page_08.txt

operantbehaviori00deitrich_Page_36.txt

operantbehaviori00deitrich_Page_38.txt

operantbehaviori00deitrich_Page_56.txt

operantbehaviori00deitrich_Page_87.txt

operantbehaviori00deitrich_Page_90.txt

operantbehaviori00deitrich_Page_33.txt

operantbehaviori00deitrich_Page_58.txt

operantbehaviori00deitrich_Page_62.txt

operantbehaviori00deitrich_Page_35.txt

operantbehaviori00deitrich_Page_65.txt


Full Text












Operant Behavior in the Human Infant:
Differentiation and Discrimination















By

SAMUEL MICHAEL DEITZ


A Dissertation Presented to the Graduate Council
of the University of Florida
In Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy


UNIVERSITY OF FLORIDA
1971






























































UNIVERSITY OF FLORIDA


3 1262 08552 3578


































This dissertation is dedicated to

Byron, Brady, Shannon, and Richard












ACKNOWLEDGEMENTS


There are many people this writer needs to thank for giving

of their time and effort toward the successful completion of a

dissertation. However, to list and publicly thank them all is

impossible. A few deserve special consideration and this writer

would like to acknowledge them here.

Special appreciation is due to this writer's advisory committee.

Dr. John M. Newell, as chairman, contributed many hours and

much advice which is highly valued. Dr. R. E. Jester, Dr. William

B. Ware, and Dr. I-. S. Pennypacker also contributed to the comple-

tion of this dissertation. All are appreciated for their advice and,

most of all, their friendship.

The parents of the infants studied in these experiments were of

great help. Dr. and Mrs. R. E. Jester, Dr. and Mrs. Barry

Guinagh, Mr. and -Mrs. William Graves, and Mr. and Mrs. John

Todd all extended much help in gathering the data for These studies.

Finally, the writer e extends his gratitude to his parents and

sisters for their support and faith in his ability.













TABLE OF CONTENTS



Page

ACKNOWLEDGEMENTS .......................... iii

LIST OF TABLES ................................. vi

LIST OF FIGURES ................................ vii

ABSTRACT .. ................................... viii


CHAPTER

ONE INTRODUCTION ................. ..... 1

TWO REVIEW OF THE LITERATURE ....... 8

Introduction ............... ...... 8
Consumable Reinforcers .......... 9
Social Reinforcers ................ 12
Tactual Reinforcers ............... 16
Auditory and Visual Reinforcers ... 17
Sum m ary ........................ 23

THREE DESIGN AND INSTRUMENTATION ..... 25

Introduction ....................... 25
Apparatus................... ....... 26
Data Collection ................... 29
Procedures ...................... 31
Experim ent I................. 32
Experiment II ................ 33
Experiment III ............... 34
Experim ent IV.................. 35
Sum m ary ......................... 36








Page


CHAPTER

FOUR RESULTS ................. .......... .. 38

Introduction ............... ......... 38
Experiment I......................... 39
Experim ent II ...................... 44
Experiment III ................... 49
Experiment IV .................... 53
Summ ary ...... ................... 58

FIVE DISCUSSION AND CONCLUSIONS ....... 59

Discussion....................... 59
Conclusions ....................... 64



APPENDIX ..................................... 70

Forms used for data collection..... 70

REFERENCES .................................... 74

BIOGRAPHICAL SKETCH.......................... 79












LIST OF#TABLES


Page




TABLE

1 Procedures Summary for Experiments I
Through IV---- ------------------ 37

2 Adjusted and Unadjusted Mean Rates for
Each Phase: Experiment I -------- 42

3 Adjusted and Unadjusted Mean Rates for
Each Phase: Experiment II ------- 47

4 Adjusted and Unadjusted Mean Rates for
Each Phase: Experiment III ------ 51

5 Adjusted and Unadjusted Mean Rates for
Each Phase: Experiment IV ------ 55













LIST OF FIGURES

Page

Figure

1. Infant leg-movement apparatus --------- ------ 27

2. Baseline and experimental phase where
right leg-movement reinforced:
Experiment I --------------------- ----- 40

3. Baseline and experimental phase where
left leg-movement reinforced:
Experiment I ----------------------------- 41

4. Baseline, extinction, and experimental phase
where right leg-movement reinforced:
Experiment II -------------------------- 45

5. Baseline, extinction, and experimental phase
where left leg-movement reinforced:
Experiment II -------------------------- 46

6. All baseline, experimental, and extinction
phases: Experiment III ------------------ 50

7. All baseline, experimental, and extinction
phases: Experiment IV------------------ 54







Abstract of Dissertation Presented to the Graduate Council
of the University of Florida
In Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy

OPERANT BEHAVIOR IN THE HUMAN INFANT:
DIFFERENTIATION AND DISCRIMINATION

By

Samuel Michael Deitz

August, 1971

Chairman: Dr. John M. Newell
Major Department: Foundations of Education

Four separate experiments were completed examining operant

response differentiation and stimulus discrimination in human infants.

The purpose of the studies was to determine whether or not infants

could learn these tasks and under what circumstances were they

most easily learned. Each experiment used one infant between the

ages of six and fourteen weeks of age as its subject, the same basic

apparatus, and visual stimuli as reinforcers.

The response studied was left and right leg-movements which

were defined as tense-extend motions of either leg which covered

a distance of at least one inch. Left and right leg-movements were

measured independently by the apparatus employed in these studies.

This apparatus also controlled the presentation or removal of the

visual reinforcers and the discriminative stimuli. Visual reinforce-

ment consisted of either light flashes or movement of a reflecting







mobile on a conjugate reinforcement schedule; the discriminative

stimuli were both visual and audible.

Baseline, experimental, and extinction phases occurred in all

studies except Experiment I where final extinction data were

unobtainable. The experimental phases differed in each study.

Experiment I examined differentiated responding under a stimulus

discrimination paradigm in one male infant using conjugate mobile

reinforcement and audible discriminative stimuli. Experiment II

was identical to Experiment I except visual discriminative stimuli

were substituted for audible stimuli. Experiment III examined re-

sponse differentiation in one female infant using light flashes as

the reinforcement. A conditioning, extinction, reconditioning

design was used in this experiment. Experiment IV was identical

to Experiment III except that one male infant was used and the

light flash reinforcement was increased in intensity.

The results of these studies showed conditioning developed in

half of the eight experimental conditions and was approached in

two others. Appropriate discriminative responding developed

under both conditions in Experiment I but developed under neither

condition in Experiment II. In Experiment II, however, one condi-

tion approached appropriate responding. Differentiated responding

developed under only one of the two conditions in both Experiments III

and IV.






It was found that the leg- movement preference demonstrated

during baseline was an important determinant of appropriate

responding in all experiments. Enhancement of that preference

effected the only conditioning in the differentiation studies. In

the discrimination studies reversal of that preference effected

more appropriate responding. In Experiment II, it was under this

reversal contingency that conditioning was approached.

The conclusions drawn from these studies included comments

on potential social and medical benefits as well as statements

concerned with future research on infant learning. It was mentioned

that one important problem for researchers of infant learning is

the discovery of reliable and powerful reinforcers. Visual stimuli

seem to lack this effectiveness and the class of consumable rein-

forcers was suggested as the best possibility. An important potential

benefit discussed was that contingency experiences such as those

incorporated in these studies may be of significant aid in building

young infants' adaptive behavioral abilities. It was also suggested

that investigation into left-right discrimination, in light of medical

evidence currently available, may provide for early identification,

prevention, and cure of behavioral deficits. Finally, it was concluded

that, in addition to these potential benefits, investigations into

infant learning can help build a complete theory of human development.












CHAPTER ONE

INTRODUCTION


Operant conditioning techniques have been effective in the

investigation of early infant learning. By discovering a stable re-

inforcing stimulus and contingently applying that stimulus as a

consequence to certain infant responses, changes in the rates of

those responses can be obtained and viewed as early, but distinct,

forms of learning. After the effective control of response rates has

been established through reinforcement, more complex forms of

learning can be investigated through the differential application of

that reinforcement. Forms of complex learning of special interest

to infant studies have included response differentiation and stimulus

discrimination.

In designing an operant experiment using infants as subjects, two

considerations are of primary importance. The first is the selection

of the response. For best results, the response must be easily

repeatable and occur naturally in the infant's repertoire. Second,

it is necessary to use an effective reinforcing stimulus which can be

made contingent upon the selected response. Having decided upon the

response and the reinforcing stimulus, contingent presentation of









the reinforcing stimulus should permit investigation into the acqui-

sition of both simple and complex behaviors. Without an effective

reinforcing stimulus, these types of investigations are impossible

as it is necessary to establish control over the response through

reinforcement in order to shape behavior through the use of dif-

ferential reinforcement.

The investigation of more complex learning through the use of

differential reinforcement demands consideration of both response

differentiation and stimulus discrimination. Although both dif-

ferentiation and discrimination involve responding, discrimination

may be a more complex form of learning as it involves responding

to specific discriminative stimuli. Both are forms of learning and

both involve a change in responding due to contingencies of

reinforcement.

Response differentiation involves responding only- This form

of learning can be defined as "making one response instead of

another" (Barrett & Lindsley, 1962, p. 424). When one response

is emitted instead of another it is because that response made is

the one that has been more frequently reinforced. Response dif-

ferentiation can be thought of as one step more complex than simple

acquisition, for acquisition only requires an increase in rate of an

undifferentiated response. In the studies reported here, response

differentiation was demonstrated by differentially reinforcing either









the left or the right leg-movements of young infants and observing

appropriate changes in the rates of these leg-movements.

Stimulus discrimination also involves responding but discrim-

ination deals primarily with differentiated responding in terms of

external, preceding, discriminative stimuli. Because one extra

variable has been added to the situation, discrimination incorporates

differentiation and can be seen as a more complex form of learning

than either simple acquisition or differentiation. Discrimination

is defined as "responding to one stimulus instead of another"

(Barrett & Lindsley, 1962, p. 424). In this form of learning, if

reinforcement is made contingent upon a response in the presence

of a specific discriminativee) stimulus, responding will increase in

the presence of that stimulus and decrease in its absence. In the

studies reported here, stimulus discrimination was demonstrated

by reinforcing one leg's movement in the presence of one stimulus

and reinforcing the other leg's movement in the presence of another

stimulus. Appropriate changes in response rates were again used

to signify reliable discriminative acquisition.

Barrett and Lindsley's (1962) findings concerning operant dis-

crimination and differentiation with retardates have implications for

research on infant learning. It was demonstrated that deficits in

the acquisition of operant discrimination and differentiation may be

used to make judgements on the capabilities of retarded children.









Since discrimination and differentiation are two component behaviors

of complex tasks, they may be used to clarify the question of

behavior deficits. Barrett and Lindsley found different patterns

of responding on tasks demanding differentiation and discrimination

with children having different types of retardation and suggested

that further research along these lines would lead to clearer infor-

mation concerning the capabilities of these children.

The studies presented here give preliminary information con-

cerning response differentiation and stimulus discrimination

capabilities in young infants. These studies are similar to those

of Barrett and Lindsley (1962) in that the same types of information

were sought. Further investigations along these lines could clarify

many questions concerning infant learning capabilities and possibly

lead to early identification of behavioral deficits.

All of the studies presented here employed the same response,

left and right leg-movement, and all of the studies used some form

of visual reinforcement. The tense-extend movement of the leg

was chosen as the response for investigation for it is already in

the infant's response repertoire; it occurs naturally at above a

zero base rate; and it can be easily repeated without fatigue As

Sheppard (1969) has commented,

The leg kick operant in infants has a great deal to
recommend it. The response is not subject to
multiple sources of control; infants can perform it









rapidly with little effort, and no apparent fatigue;
the operant level is above zero and the maximum
response rate is quite high; and a transducer for
the response is extremely simple and easily
constructed. Given all these characteristics,
it would be an excellent response to utilize in
future studies (p. 48).

Sheppard's reasoning was helpful in choosing the leg-movement

response for these investigations.

Visual stimuli were chosen as the reinforcers because they

have been shown to be effective in the past by a number of studies,

and because they are easily adaptable to automatic presentation.

Visual stimuli can also be easily varied in intensity and duration

thus making comparative efficacy studies feasible. A more com-

plete discussion of these supporting studies and others is pre-

sented in Chapter Two.

Two types of visual reinforcement were used in these investi-

gations. In the differentiation studies, light flashes of varying

intensity were made contingent upon the appropriate leg-movements.

This is a conventional method of reinforcement employing a discon-

tinuous reinforcing event. In the discrimination studies, the less

conventional method of conjugate reinforcement was employed

using the continuous presence of a mobile, where movement of

the mobile was contingent upon leg-movement.





6


Conjugate reinforcement (Lindsley, 1957, 1961, 1962; Lindsley,

Hobika, & Etsten, 1961; Rovee & Rovee, 1969) is a reinforcement

system where "the intensity of a continuously available reinforcing

stimulus varies directly and immediately with the rate of response"

(Lindsley, 1962, p. 3). In these discrimination studies, movement

of the mobile, contingent upon leg-movement, rather than its mere

presence served as the reinforcing stimulus. Rapid leg-movement

produced rapid movement of the mobile. Thus, the reinforcing pro-

perty of this stimulus was found in its intensity, and the intensity

varied directly with leg-movement.

Lipsitt (1963), among others, has recognized the importance of

investigating learning abilities of the young infant and has mentioned

the lack of current research in that area. "A review of the available

data concerning the learning process of the very young child reveals

that comparatively few studies are available, and few of these have

systematically explored the various parameters known to affect

learning in lower animals or the older person" (Lipsitt, 1963,

p. 148). Some aspects of those parameters were investigated in

these studies; those were the parameters dealing with infant learning

of response differentiation and stimulus discrimination through

reinforcement.

It is the intention of the studies presented here to demonstrate

that it is possible for infants to learn complex tasks. Specifically,






7



they are presented to give preliminary data on the acquisition of

left-right leg-movement discrimination and differentiation by human

infants. Evidence is also given showing that visual stimuli can

function as reinforcers for establishing these complex behaviors.

The purpose of these studies is to establish discriminative and

differentiated responding in human infants using visual stimuli as

the reinforcement.















CHAPTER TWO

REVIEW OF THE LITERATURE


Introduction


To further the investigation of infant behavior through the use

of operant conditioning techniques, it is necessary to identify stimuli

that have been used as positive reinforcers and have been shown to be

both powerful and stable. It is also important to consider the types

of responses that have been studied with these reinforcers. This

review of the literature serves to classify the types of reinforcers

that have been used in the various infant studies and to point out the

type as well as the level of complexity of the response involved.

This review is limited to only those studies using operant conditioning

techniques with infants from birth to the first year of life.

Many articles have reviewed studies dealing with operant

behavior in the first year of the infant's life (Lipsitt, 1963, 1966,

1967; Bijou & Baer, 1966), but only two have presented a plan for

grouping reinforcing stimuli into different categories of a classifi-

cation system (Bijou & Sturges, 1959; Hulsebus, 1970). The benefits

of having a common system of classification of reinforcing stimuli









are many, but primarily that system allows for easy access to

information concerning the effectiveness of the different types of

reinforcers. The systems developed by Bijou and Sturges (1959)

and Hulsebus (1970) do not perfectly adapt in all categories to

classifying information on very young infants, so a modification of

their categories is used here. For this review, reinforcers are

categorized into the following classes: consumable reinforcers,

social reinforcers, tactual reinforcers, and auditory and visual

reinforcers. Within each category, responses are discussed

according to the following proposed levels of complexity. First,

those studies dealing only with acquisition are discussed, followed

by studies dealing with differentiation and finally studies establishing

discriminative repertoires.


Consumable Reinforcers

A wide variety of consumable reinforcers has been used in

operant investigations with infants. Honey, jam, saccharine,

M & M's, bits of food, sugar, dextrose solution, and milk have all

proven to be effective reinforcers for infants. The great majority

of the studies, however, employed the last two solutions delivered

either by nipple or tube. Lipsitt, Kaye, and Bosack (1966) conditioned

the sucking response of infants under four days of age. A significant

increase over baseline rates was shown in sucking a 1/4 inch rubber









tube by using the contingent presentation of a 5 percent dextrose

solution for five seconds. Sucking rates increased from baseline

during conditioning and decreased during extinction. There was no

increase in rate over baseline during the same period for a control

group.

Siqueland (1964) used milk delivered via a standard nipple to

demonstrate differentiation in three-month-old infants. Reliable

acquisition and extinction of either right or left head turns by

differentially reinforcing the response only when made in the

desired direction was demonstrated. No change was found in

either direction for a control group.

Other studies have demonstrated some type of discrimination

training using consumable reinforcers with infants. Siqueland and

Lipsitt (1966) used infants under four days of age to demonstrate

discrimination learning. A 5 percent dextrose solution was used

as the reinforcing stimulus and satiation effects were controlled

by keeping the time since feeding constant for all subjects. Rein-

forcement of the response, ipsilateral head rotation, was accom-

plished by a two second presentation of the solution via a standard

nipple. In the beginning of the experiment, reinforced head turns

were cuedd" by a touch to the cheek and a tone to the side indicating

reinforcement. The authors then proceeded to discrimination

training; responses in the presence of one tone were reinforced while

responses in the presence of another were not. A reversal was









demonstrated and reliable acquisition shown for all conditions.

Papousek (1967) also demonstrated discriminative head turn

behavior in infants as young as three months. Left head turns were

reinforced with milk in the presence of a bell and right head turns

were reinforced in the presence of a buzzer. Acquisition of the

original discrimination and a reversal were demonstrated.

In other studies the same types of procedures were used. Valen-

tine (1914) established color discrimination in four- and eight-month-

old infants using honey and jam as the reinforcers. Ling (1941)

investigated form discrimination in infants as young as six months

and achieved reliable results. Saccharine was used as the reinforcer

in his studies. Rey and Rey-Pinto (1959) established color discrimi-

nation in a ten-month-old infant by reinforcing the correct choice with

sugar. Hill (1965) used M & M's and bits of food to establish object

discrimination in infants as young as twelve months.

Consunable reinforcers have proven to be quite effective in

establishing operant behavior in infants. Although only one study

found (Siqueland & Lipsitt, 1966) formally controlled for satiation

effects, the results obtained by the various studies seem to indicate

that this is not a problem in infant studies. Complex discriminative

responses as well as simple acquisition have been established using

consumable reinforcers and these reinforcers, as a class, seem to

be highly effective and enduring. There is much promise for future

studies employing the use of consumable reinforcers.








Social Reinforcers

Most human learning seems to take place under the control of

social stimuli. Consequently, many studies of infant behavior have

used social stimulation as the reinforcing agent. Different types of

human verbal responses, the presence or absence of a human experi-

menter, bodily contact, and facial expressions have all been used as

reinforcers in some of the studies reported on early infancy. As a

category, social reinforcement contains innumerable specific stimuli,

but the following studies demonstrate a beginning toward identifying

those stimuli which seem to be the most important.

Three studies in the literature have built upon each other in an

attempt to separate those factors in social reinforcement which are,

in fact, reinforcing. Rheingold, Gewirtz, and Ross (1959) used social /

reinforcement to condition vocal behavior in three-month-old infants.

After two days of baseline, the conditioning phase consisted of an

experimenter leaning over the crib, expressionless until the infant

produced a discrete vocal sound. The experimenter would then change

his expression briefly and return to the expressionless condition until

another sound was produced by the infant. This phase was followed by

two days of extinction. Reliable acquisition and extinction were demon-

strated and the experimenters concluded that a "social event composed

of an everyday complex of acts, performed by an adult who is not the

caretaker, can function as a reinforcing stimulus" (p. 72).

These results have been questioned since the presence of an
immobile adult may serve as a discriminative stimulus for vocal







behavior in the infant or as a type of "social release. Weisburg (1963)

controlled for this possibility in an experiment which stemmed from

the above results. The subjects in this experiment were also three

months old and the response definition for vocal behavior was the

same. He divided his subjects into the following six groups. Group

one had no experimenter present. Group two had a constantly immobile

experimenter present with a "blank expression" at all times. Group

three received noncontingent social reinforcement while group four

received noncontingent, nonsocial reinforcement consisting of a door

chime. Group five received contingent social reinforcement and group

six received contingent nonsocial reinforcement. Vocalizations increased

over baseline only in the group receiving contingent social reinforcement;

thus, this condition is the only one where an actual reinforcer was found

to be acting. A functional definition of reinforcement disqualifies the

other stimulus consequences being called reinforcers. Weisburg con-

cluded that "the results indicated that the initial presentation of an

unresponding human did not serve as a discriminative stimulus (SD)

for vocal behavior" (p. 386).

Todd and Palmer (1968) further clarified this issue by again using

three-month-old infants and the vocalization response. One of

the groups received contingent tape-recorded verbal stimulation

with an adult present (AP) and the second group received the same

stimulation with no adult present (NAP). By using a sound activated

relay system, vocalizations produced the tape-recording for five





14


seconds. There was no significant difference between the groups

during either baseline or extinction, but the AP group increased

significantly higher than the NAP group during conditioning. Still,

both groups increased significantly over baseline during conditioning.

The results indicate that the "primary influence on infant babbling

is the human voice" (p. 595). Further clarification of this issue

concerning which factors of a social act are, in fact, reinforcing

will come from future research.

Two other studies were concerned with the comparative efficacy

-of different types of social reinforcers. Koch (1967) used fifty infants

at the end of their second month and compared the strength of five

different social consequences on the head rotation response. The

three of his five consequences which involved changing stimuli worked

the best and he concluded that the strongest reinforcer is a new and

changing stimulus. In comparison to stimulus variety, it was found

that the mother's voice and face are comparatively weak reinforcers.

Brossard and Decarie (1968) compared eight different social -

consequences while working with the smiling response of infants

between four and five months of age. The following stimuli were used

in their experiment and are listed from the most to the least effective:

picking up the infant, smiling, rocking, picking up and speaking,

touching and speaking, speaking, and touching.









Other studies have dealt with simple acquisition and social

reinforcement. Charlesworth (1966) conditioned orienting and

attending behavior in five- to ten-month-old infants using social

reinforcement. The social reinforcement consisted of a tape-

recorded human voice saying "ah . peek-a-boo" and the presen-

tation of a terry cloth face. There was significant increase in response

rate over the baseline period during conditioning and a return to baseline

through extinction.

Brackbill (1958) has been the only researcher to report the

effects of certain reinforcement schedules on extinction data. The

smiling response of three-and one-half- to four-and one-half-month-

old infants was recorded; the reinforcing stimuli used social and

bodily contact. One group received continuous reinforcement (CRF)

during the entire conditioning phase of the experiment. The second

group received continuous reinforcement at the beginning of the phase

but was soon switched to fixed ratio schedules of values two, three

and four. The group receiving the intermittent reinforcement was

found to be more resistant to extinction and Brackbill concluded that

"intermittent reinforcement is superior in maintaining continuous

performance of a response during extinction" (p. 124). This result

has long been established in the operant literature using animals as

subjects but Brackbill's study was the first, and to the present, only,

to report these findings with human infants.









All of the studies reported above deal with simple response

patterns. Although no studies using social reinforcement have been

reported which establish discriminative behavior, Routh (1969) con-

ditioned differentiated vocalizing using social reinforcement. Thirty

infants between the ages of two and seven months old were divided

into three groups. The reinforcement for each group consisted of a

smile, three "tsk" sounds, and a light stroke of the infant's abdomen.

One group was reinforced only for consonant sounds, a second group

only for vowel sounds, and a third group for any vocalization. All

groups increased from baseline, and significant response differentiation

occurred in the predicted direction. It was concluded that the results

were consistent with a conditioning approach to beginning language

learning.

Social reinforcement presents the intricate problem of deter-

mining what aspects of a social act are reinforcing. However, from

the above literature, it is evident that this class of reinforcers is a

viable one with which to work. If social stimulation is to be shown as

responsible for much of our adult behavior, more studies need to be

done along the lines of that by Routh (1969) where the beginnings of

complex repertoires were established and maintained.


Tactual Reinforcers

Few studies reported in the literature used tactual stimuli as

the reinforcing consequences of infant behavior. Schaefer (1960)

conditioned button pushing in a seven-month-old infant by applying a









soft, vibrating electric toothbrush to the right palm of the infant for

ten seconds contingent upon each button push. The rate of button

pushing increased from baseline during conditioning and decreased

during extinction. Sheppard (1969), in a study covered fully under

Auditory and Visual Reinforcers, conditioned vocalizations and leg

kicking using a similar procedure.

In the study by Brossard and Decarie (1968) discussed earlier,

touching of the infant was found to be the least effective method of

reinforcement among the eight different consequences tried. Touch-

ing as defined by those authors, however, was simply the application

of the palm of the hand to the infant's abdomen; no tickling or vibration

was involved. The two studies discussed above indicate that tactual

reinforcement can be effective if it involves movement of the tactual

stimulus, such as vibration; it is likely that an immobile tactual

stimulus like that employed by Brossard and Decarie (1968) is an

ineffective reinforcer for infant behavior.


Auditory and Visual Reinforcers

Auditory and visual reinforcers have been used extensively in

infant conditioning studies for a number of reasons. Their intensity

and duration can be controlled; they are easily presented by automatic

equipment; and they have been found to be effective over a wide range

of responses. Such varied stimuli as tones, mobiles, lights, patterns,








slides, and chimes have all been shown to act as reinforcers for

infants under one year of age. These stimuli cannot be thought of

as primary reinforcers nor have they been systematically paired

with primary reinforcers to insure an acquired reinforcing effect.

Their reinforcing ability probably occurs because of the effective-

ness of what Bijou and Baer (1965) called "general stimulus change."

Simple acquisition has been demonstrated using normal and

conjugate reinforcement in infants under one year of age. Rheingold,

Stanley, and Cooley (1962) conditioned touching in four-month-old

infants by the contingent presentation of colored geometric forms

for one and one-half seconds following each touching response. Rovee

and Rovee (1969) demonstrated the effectiveness of conjugate reinforce-

ment of the leg-movement response of ten-week-old infants. An L-

shaped mobile was always in the infant's line of vision but the infant

could move the mobile and produce loud clicks by rapid leg-movement.

The experimental group received contingent conjugate reinforcement

while the control group received noncontingent stimulation. The leg

kick rates of the experimental group tripled over baseline while the

rates of the control group showed no significant change.

Watson (1970) also used a mobile over the head of his eight-

week-old infants. The head press response was employed and each

press earned a one second rotation of the mobile. The experimental

group showed a significant increase in rate while both control groups,






19

one receiving noncontingent stimulation and the other viewing a stable

mobile, showed no increase. Lipsitt, Pederson, and DeLucia (1966)

used a conjugate reinforcement system where the light illuminating

a viewing box varied in intensity according to the rapidity of lever

pressing. Lever pressing rates in twelve-month-old infants increased

from baseline during both conditioning and reconditioning and decreased

during extinction. Smith, Johnson, and Lindsley (1967) conditioned a

three-month-old infant to look at another person's eyes by contingently

applying the multiple stimulus of a smile and a hand clap on each change

in "looks. Responding increased from baseline during conditioning

and decreased again during extinction.

Three studies have demonstrated differentiated infant responding

using auditory and visual reinforcers. Watson (1967) used both auditory

and visual reinforcement to alter the direction of visual fixation in

thirteen- and fourteen-week-old infants. In two experiments, one with

twenty-seven infants and the other with twenty-four younger infants,

a significant increase in differential spatial fixation was obtained during

the conditioning phases as compared with the baseline phases.

Caron (1967) used a broad range of visual stimuli as reinforcers

while working with the head rotation response in three- and one-half-

month-old infants. An increase over baseline was demonstrated, fol-

lowed by extinction and reconditioning of the response. Differen-

tiation was demonstrated with some subjects by conditioning a reversal

from right head rotation to left head rotation. These results lend status





20 *


to sensory events as reinforcers and, as she summarized, "this

research furnished additional evidence that exteroceptive feedback

can promote and sustain behavior in infancy" (p. 509). Levison and

Levison (1967) also demonstrated differential conditioning of head

turning in three-month-old infants. The visual reinforcement con-

sisted of a variety of slides projected on a screen. They conditioned

head turning to one side,then after an extinction period reversed the

head turning to the other side. Agreeing with Caron (1967), Levison

and Levison (1967) conclude that, "rapid conditioning of head turning,

extinction and conditioning of a response reversal can be obtained in

human infants using visual stimulus patterns as reinforcing stimuli"

(p. 530).

Three studies have reported failure, or partial failure, in

conditioning infant operants using auditory and visual reinforcers.

Friedlander (1961) failed to obtain evidence of simple acquisition of

cord pulling in two infants, three and seven months old respectively,

using any of four visual stimuli as possible reinforcers. Lipsitt (1963)

improved Friedlander's apparatus in an attempt to condition discrimi-

native behavior in eight-month-old infants but did not demonstrate

acquisition of that behavior.

Watson (1969) obtained mixed results in his attempt to condition

visual fixation in infants from ten to fourteen weeks of age. Visual










fixation was differentially reinforced with visual and auditory stimuli.

In the first study, using fourteen-week-old infants, acquisition was

found for female infants under auditory stimulation but not under

visual stimulation; the opposite was true for male infants. Using

ten-week-old infants in a second experiment, consistent learning was

found for girls but no significant learning for boys under either, or a

combination of both, types of stimulation.

The studies involving discrimination learning in infants using

auditory and visual reinforcement have been quite successful. Leuba

and Friedlander (1968) conditioned the discrimination between lighted

and unlighted toys in seven- to eleven-month-old infants. When the

lighted toy was touched, door chimes would sound and varied colored

lights would flash. The unlighted toy produced neither auditory nor

visual feedback. All twenty infants manipulated the lighted toy more

often; it was pressed twice as often and for three times as long as

the unlighted toy.

Another discriminative response was established in twelve-month-

old infants by Simrrmons (1964). Using "tapping" behavior as the response

and the sound of door chimes as the reinforcing stimulus, acquisition

of a red-blue discrimination was demonstrated. The infants were

divided into two groups and placed before a panel with two circles on

it; these circles could be illuminated by either a red or a blue light.





22 *


For one group, the red circle signaled reinforcement and the blue

signaled extinction. In the other group the program was reversed.

The location of the color on the panel, left or right, changed every

minute for both groups to control for position effects. Discriminative

behavior developed for both groups during conditioning and disappeared

during extinction.

Sheppard (1969) conditioned two concurrent operants, leg

kicking and vocalizations, in an attempt to achieve differential respond-

ing under discriminative control in one infant less than three months

old. Both operants were placed successfully under separate schedule

control using the mother's voice and flashing red and yellow lights as

the reinforcing stimuli. At that point, two discriminative stimuli

were introduced and responding was conditioned to occur under

stimulus control. During this experiment, both differential responding

and discriminative responding were achieved,and to the present, this

experiment is the best demonstration of pure operant research in

infants.

With just a few exceptions, auditory and visual reinforcers have

been shown to be highly effective in controlling infant operant behavior.

Because of their flexibility and variety, and because they are easily

controlled and presented automatically, auditory and visual stimuli

are quite adaptable to operant conditioning experiments. Because of

these reasons, reinforcers of this type have been used in the experi-

ments presented in this report.










Summary

A review of operant conditioning studies on infants in their

first year of life demonstrates that operant conditioning is a research

technique of considerable importance towards understanding learning

in the human infant. It further justifies the statement made by Lipsitt

(1967) in concluding one of his reviews on learning in the human infant.

"Current data on infant learning suggest that a renewed optimism is

warranted with respect to the learning potential of infants, as well as

to the feasibility of studying that learning potential" (p. 246).

It has been shown that a wide variety of reinforcing stimuli is

effective in controlling infant operant behavior. The behavior

investigated ranges from simple response acquisition through differen-

tiation and discrimination. Still, there are many areas where more

work should be done; specifically, questions such as the following

need be answered: Which aspects of a stimulus complex are, in fact,

reinforcing? Which types of reinforcers are the most effective? Can

developmental levels be determined from investigating the learning of

complex tasks? Can early identification of behavioral deficits aid in

overcoming those deficits?

Investigation into response differentiation and stimulus discrim-

ination can begin to answer some of the above questions. No work has

yet been done which investigates infant response capabilities with

these types of learning. The studies presented here begin to analyze,





24 #




with infants, the questions dealt with by Barrett and Lindsley (1962)

with retarded children. The results presented here are preliminary,

but suggest directions in which future research on infant learning

could move.













CHAPTER THREE

DESIGN AND INSTRUMENTATION


Introduction


Four separate experiments, each using visual reinforcement

and the same basic apparatus, were conducted to investigate differ-

entiation and discrimination in the human infant. The operant response

used for these investigations was the leg-movement response. Leg-

movement was defined as a tense-extend motion of the leg which

covered a distance in either direction of at least one inch.

Each experiment used one infant chosen from those whose parents

volunteered them for these studies. The only criteria for subject

selection were that the infants be full term births and that they be

judged medically healthy at the time the study began. The infants

were between six and eight weeks old when the studies began and had

not reached the end of their fourteenth week when the studies ended.

The unusually long time interval for these studies allowed for experi-

mental adjustments during the investigations as they were found

necessary by inspection of the data.





26 *


Apparatus

The apparatus used for this series of experiments was designed

to be adapted to all conditions in the four experiments. This section

explains the different possible uses of the apparatus, as well as its

dimensions, while the specific methodology for how each experiment

was conducted is mentioned in a following section.

The apparatus (see Figure 1) consisted of a partially enclosed

box, twenty-eight inches long, nineteen inches high and twenty-three

inches wide. The front, back, and part of the top were open to allow

for visual inspection of, and easy access to, the infant. Sides were

present to prevent the infant from receiving excess or uncontrolled

visual stimulation which could be received by the turning of the head.

All interior parts of the apparatus were painted a dull, white color to

prevent visual stimulation which might compete with the visual stimuli

used for the purposes of the studies. Two boards, each seventeen

inches long, projected from the back of the apparatus and allowed

for ribbons to be attached from the ends of the boards, around a

fulcrum, to the infant's ankles. When the infant tensed its leg the

board was pulled down; springs attached to the board pulled the board

back to its original position when the infant extended its leg.

The apparatus counted the movements of each leg separately

in either of two ways. First, mechanical counters were mounted

so that when a board was pulled down the lever on the counter was

























































Figure 1. Infant leg-movement apparatus









depressed and the movement was registered. The leg had to be

extended and tensed again in order to have another count registered.

Second, microswitches were mounted so that when the infant tensed

its leg they were depressed and the count was registered on automatic

equipment located on the back of the apparatus. A tense-extend

movement of at least one inch in both directions was required to

depress either the mechanical counters or the microswitches. Both

methods were used in these experiments, though never simultaneously.

Due to limitations in the electrical equipment, an inner response time

(IRT) of .75 seconds was demanded for either the count to register or

reinforcement to be applied when the microswitch method was being

used. This was not found to be a problem, however, due to the some-

what low rate of infant leg-movement.

Two types of visual stimuli were made automatically contingent

upon leg-movement using this apparatus. First, a mobile was attached

to either or both boards controlled by leg-movement to supply conjugate

reinforcement. The reinforcing property of the mobile was found in

the intensity of its movement rather than merely its presence; the

intensity of mobile movement was a direct result of leg-movement.

The second type of visual stimulus which was made contingent

upon infant leg-movement was a brief flash of colored light. The

lights were located underneath the board on which the counters were

mounted. Depending on the purposes of the experiment in progress,









the reinforcing stimuli were made contingent on either left or right

leg-movement, leaving movements of the other leg on an extinction

schedule.

Two types of discriminative stimuli were also available for use

with this apparatus. The first type of discriminative stimrrlus used

was audible stimuli presented through a loudspeaker located behind

the infant's head. Audible reinforcers may also have been presented

through this loudspeaker but were not used in these studies. The

second type of discriminative stimulus used was visual. These

employed red and green colored lights located above the infant's

head. These visual stimuli were used when the conjugate mobile

reinforcement was used.

The flexibility of this apparatus was utilized during the different

phases of the four experiments conducted for this study- Both types

of counting mechanisms were used as well as both types of reinforcing

stimuli, flashing lights for one type of reinforcement and a reflecting

mobile for the other. Both audible and visual discriminative stimuli

were also used in these studies, a buzzer during one experiment and

red and green lights during another.


Data Collection

Baseline, conditioning, and extinction phases occurred in all

but one experiment where extinction data were not collected. The

conditioning phases differed somewhat from experiment to experiment.










Data were obtained in each phase with the infant being placed in the

apparatus with the appropriate contingencies in effect for an experi-

mental "session. Sessions occurred at least once per day at

approximately the same times; more than one session often occurred

in the same day and sometimes a few days would lapse between sessions.

A session did not take place if any of the following conditions were

present in the infant at the time of the session: (1) sickness,

(2) sleepiness, (3) crying, (4) on drugs or medication, and (5) too

"grumpy. The length of a session depended upon the availability

of the infant and its general condition during the session. If the

infant began to cry or fall asleep during the session, the session

was terminated immediately. No session was considered represen-

tative unless it extended for at least three consecutive minutes.

The baseline phases extended for at least four sessions as did the

extinction phases. The conditioning phases extended for at least

ten sessions or longer depending on the data.

Data were collected in terms of average rate of leg-movement

per session. One or both of the infant's parents were trained to

position the infant properly in the apparatus. For each session, the

parent placed the infant in the apparatus, turned the counters to zero

at the beginning of the session, and kept the length of time of that

session by using a stopwatch provided for that purpose. When the

session was completed, the parent marked the frequency of movement









of each leg and the elapsed amount of time of the session on special

forms provided for that purpose. Rate data were computed from

the time and frequency data collected by the parents.

The forms that were used for data collection are shown in the

Appendix. All forms have columns for left and right leg-movement

frequencies, left and right leg-movement rates, and the elapsed time

of each session. Form one was used to report baseline and extinction

data for Experiments I and II; form two was used in the conditioning

phases of Experiments I and II. Form three was adaptable to all

phases of Experiments III and IV. The column designating switch

position (SP) indicates which conditions were in effect during that

phase; the switch is the one located on the electrical equipment panel

on the apparatus.

The forms were collected from the parents weekly with equip-

ment or design checks made every three or four days. The experi-

menter was available at all times for any consultation the parents

of the infant required. The experimenter controlled the length and

duration of any phases and decided on when phase changes occurred.

These decisions were made based on examination of the data.


Procedures

Four experiments were conducted, each with slightly different

goals but all investigated response differentiation or stimulus dis-

crimination in the human infant. Experiments I and II were primarily









concerned with stimulus discrimination but contained elements of

differentiation. Experiments III and IV dealt only with response

differentiation. The following explanations point out the specific

procedures used in each experiment.


Experiment I

Experiment I examined stimulus discrimination and response

differentiation in one male infant. The experiment began when the

infant was six-weeks-old and ended at his twelfth week. The res-

ponse was the leg-movement of both legs and the reinforcing stimulus

was the conjugate reinforcement of the reflecting mobile. The

apparatus was adjusted to count mechanically for this experiment.

Baseline and experimental phases were instituted but final extinction

data were unattainable because of illness to the infant.

The baseline phase consisted of obtaining rate data on the move-

ment of both legs. The mobile was not present for this phase. Base-

line data were gathered for a total of fourteen sessions. The left-

right leg-movement discrimination was attempted during the experi-

mental phase. Two conditions occurred during each experimental

session. These conditions were presented in a random manner over

all ten conditioning sessions to control for sequence effects. In one

condition, in the presence of a buzzer, the mobile was attached to

the board controlled by the left leg. Under this condition, left









leg-movement was reinforced on a continuous conjugate reinforcement

schedule. This condition placed right leg-movement on an extinction

schedule. In the second condition, a no-buzzer or silent condition,

the mobile was attached to the board controlled by the right leg.

Here, right leg-movement was on the continuous conjugate reinforce-

ment schedule and left leg-movement was on an extinction schedule.

As mentioned, no final extinction data were obtained.


Experiment II

Experiment II also investigated left-right leg-movement discrim-

inative and differentiated responding in one male infant. This study

began at the infant's sixth week and ended at his fourteenth week.

The reflecting mobile served as the conjugate reinforcement and

counting was accomplished mechanically. The same procedures were

used as in Experiment I with discrimination training occurring under

two conditions within the same session. In Experiment II, however,

the discriminative stimuli were red and green lights.

Twenty-one baseline sessions were completed where neither the

mobile nor the discriminative lights were present. One condition

used in this study's experimental phase consisted of the mobile

being placed on the board controlled by the infant's right leg in the

presence of a green light. The green light signaled continuous conjugate

reinforcement for right leg-movement and extinction for left









leg-movement. The second condition consisted of the mobile being

placed on the board controlled by the infant's left leg in the presence

of a red light. The red light signaled continuous conjugate reinforce-

ment for left leg-movement and extinction for right leg-movement.

Again the conditions were randomly alternated to control for a

sequence effect. Twenty experimental sessions were held followed

by six sessions of a final extinction phase.


Experiment III

Experiment III used one female subject to investigate differentiation

in the leg-movement response. This study began at the infant's seventh

week and ended at her fourteenth week. The reinforcer was a brief

(. 5 second) flash of red light from one bulb located above the infant's

head. The apparatus for this experiment was set up to count leg-

movement electrically by depression of the microswitches.

Baseline, two experimental, and two extinction phases occurred

in this experiment. The baseline and extinction phases were identical;

leg-movements were counted but no light flash followed any movement.

The first experimental phase consisted of thirteen sessions where left

leg-movement was reinforced on a continuous schedule of light flash.

During all thirteen sessions only left leg-movement was reinforced;

right leg-movement was thus placed on an extinction schedule. The

second experimental phase consisted of eleven sessions where only the










right leg-movement was reinforced on a continuous reinforcement

schedule for light flash. Left leg-movement during this phase was

placed on an extinction schedule. The original baseline phase consisted

of eight sessions; the first extinction phase consisted of four sessions;

and the final extinction phase consisted of four sessions. No discrim-

inative stimuli were used during this experiment since response

differentiation rather than stimulus discrimination was the object

of investigation.


Experiment IV

Experiment IV also investigated response differentiation in one

male infant. It was begun at the infant's sixth week and ended at his

twelfth week. The reinforcing stimulus was again a brief flash of

light, except, for this experiment, three lights flashed simultaneously.

One light remained red; the others were amber and yellow. Counting

was again accomplished electrically.

Baseline, experimental, and extinction phases were used in

Experiment IV as they were in Experiment III. The only change

consisted of the right leg-movement being reinforced in the first

experimental phase instead of the left leg-movement. Left leg-movement

was reinforced during the second experimental phase; this was a

reversal of the sequence which occurred in Experiment III. Baseline

consisted of five sessions, the first experimental phase of ten sessions,










extinction of five sessions, the second experimental phase of

eleven sessions, and final extinction of five sessions. Again no

discriminative stimuli were used since none were called for by

the purposes of this study.


Summary

Four studies were completed to investigate differentiated and

discriminative responding in the human infant. Two studies dealt

primarily with stimulus discrimination although they contained

elements of differentiation, and two with response differentiation.

The same apparatus was used for all four studies. Table I summarizes

these four studies and describes the phases which occurred in each.






TABLE 1
Procedures Summary for Experiments I Through IV

RESPONSE: Tense-extend movements of the right and left legs
Experiment I
Objective: Discrimination and Differentiation
Reinforcement: Conjugate mobile, CRF
Phases:
1. Baseline
2. Experimental
A. Buzzer Left leg-movements reinforced
Right leg-movements extinguished
B. No-Buzzer Right leg-movements reinforced
Left leg-movements extinguished
Experiment II
Objective: Discrimination and Differentiation
Reinforcement: Conjugate mobile, CRF
Phases:
1. Baseline
2. Experimental
A. Red light Left leg-movements reinforced
Right leg-movements extinguished
B. Green light Right leg-movements reinforced
Left leg-movements extinguished
3. Extinction
Experiment III
Objective: Differentiation
Reinforcement: One light flash, CRF
Phases:
1. Baseline
2. Experimental I
Left leg-movements reinforced
Right leg-movements extinguished
3. Extinction I
4. Experimental II
Right leg-movements reinforced
Left leg-movements extinguished
5. Extinction II
Experiment IV
Objective: Differentiation
Reinforcement: Larger intensity light flash, CRF
Phases:
1. Baseline
2. Experimental I
Right leg-movements reinforced
Left leg-movements extinguished
3. Extinction I
4. Experimental II
Left leg-movements reinforced
Right leg-movements extinguished
5. Extinction II














CHAPTER FOUR

RESULTS


Introduction



The four experiments presented in this study were analyzed

separately but similar statistical procedures were employed for

each. The data from the individual studies are plotted on semi-log

graph paper and described in the text. Mean leg-movement rates

were computed for the different phases of the experiments from the

average rates of leg-movement per session and are presented in

tables accompanying the text.

The data were analyzed using the "t"-test for related samples

(Edwards, 1969, pp. 135-137) after the experimental sessions were

adjusted for initial baseline differences. The adjustments were

made by subtracting the mean baseline rate of the leg-movement

under consideration from each of its experimental session rates.

Adjusted mean experimental rates were then computed and compared

by the use of the "t"-test. Significant differences in the expected

direction indicated successful conditioning during that phase.








Experiment I

The results of this study are shown in Figures 2 and 3. Figure 2

contains the fourteen sessions of baseline data and the ten sessions

of data concerning that part of each experimental session when the

buzzer was off and right leg-movements were reinforced. Figure 3

illustrates the same baseline data and the ten sessions of data con-

cerning that part of each experimental session when the buzzer was

on and left leg-movement received reinforcement. The different

conditions in the experimental phase are presented separately to

better illustrate how each condition compares to baseline. Table 2

shows both the adjusted and the unadjusted means for each phase

of this experiment.

The graphs illustrate effective conditioning under both conditions.

During the no-buzzer component, when right leg-movement resulted

in reinforcement, both right and left leg-movement rates increased

from baseline, but rightleg-movement remained more frequent

than left leg-movement. A reversal can be noted by observing

Figure 3; left leg-movement rates exceeded right leg-movement rates

during the buzzer component.

Table 2 also demonstrates that discriminative behavior developed

under both conditions. During the no-buzzer component, the mean

right leg-movement rate increased from a baseline of 2.'98 to an

experimental mean of 9. 09. Although unreinforced in this component,





1000
500



100
50 Baseline Experimental
1SD Right

10






- e i ;
41 5
1 1
5 F





.05


.01 Right Leg-Movement

.005 ...... 'Left Leg-Movement
(---x Record Floor -
Reciprocal of Session
.001 Length
10 20 30 40 50 60
CONSECUTIVE SESSIONS
Figure 2. Baseline and experimental phase where right leg-movement
reinforced: Experiment. I






1000

500



100

50



10
5



1

.5



.1

.05



.01

.005



.001


10 20 30 40 50 60
CONSECUTIVE SESSIONS
Figure 3. Baseline and experimental phase where left leg-movement
reinforced: Experiment. I








TABLE 2


Adjusted and Unadjusted Mean Rates for Each Phase: Experiment I


Leg-Movement Baseline Experimental
(Unadjusted) (Unadjusted)
Buzzer (SD Left) No-Buzzer (SD Right)

Right 2.98* 4.62 9.09


Left 2. 14 9. 95 5. 05

Leg-Movement Baseline Experimental
(Adjusted) (Adjusted)
Buzzer (SD- Left) No-Buzzer (SD Right)

Right 0.00 1.64 6.12***


Left 0.00 7. 80** 2.91


)*p 20.
**p4 01.
***pL. 001.









left leg-movement rates also increased but from a baseline mean

of 2. 14 to an experimental mean of 5. 05. Still, left leg-movement

rates remained below right leg-movement rates as would be expected.

During the buzzer component, the mean left leg-movement rate increased

from its baseline of 2. 14 to an experimental mean of 9. 95. Right leg-

movement rates also increased during this component from a baseline

of 2. 98 to an experimental mean of 4. 62. As expected, the right leg-

movement rates remained below the left leg-movement rates in this

component.

Although the initial baseline rates showed a right over left preference,

there were no statistically significant differences between their means

(t=l. 36, p4.20). Also, there were six sessions during baseline where

left leg-movement exceeded right leg-movement. Analysis of the

adjusted experimental means supports the contention that reliable

discriminative responding developed. During the no-buzzer condition,

right leg-movement rates were significantly higher than left leg-

movement rates (t=3. 28, p/. 01). There was only one session where

left leg-movement exceeded right leg-movement in this component.

During the buzzer condition, left leg-movement rates were significantly

higher than right leg-movement rates (t=6. 35, p. 001), and there were

no sessions where right leg-movement exceeded left leg-movement.





44


Experiment II

The outcome of this study is presented in Figures 4 and 5.

Figure 4 presents the twenty-one session baseline, six session final

extinction, and also the data from that part of each experimental

session when the green light was on and right leg-movements reinforced.

Figure 5 presents the same baseline and extinction data and illustrates

twenty sessions of the data concerning that part of each experimental

session when the red light was on and left leg-movements were

reinforced. Table 3 shows the adjusted and unadjusted means for

each phase of this experiment.

Observation of the graph shows differences between leg-movement

rates during both baseline and the red light experimental component.

There is a left leg-movement preference during baseline which is

somewhat overcome during the green light experimental component

where right leg-movements are reinforced. During the red light

experimental condition, that initial baseline difference is expanded.

Extinction data are most similar to that of the first four baseline

sessions.

During baseline, the left leg-movement mean rate exceeded the

right leg-movement mean rate; the differences were significant

(t=3. 14, p/. 01). There were six sessions where right leg-movement

surpassed left leg-movement. In the red light experimental condition,

left leg-movement increased slightly from a baseline mean of 7.43





1000

500



100

50


--\ ,/ \ \/ '
K ,


Experimental
SD Right


Extinction


Right Leg-Movement
-----. Left Leg-Movement

Record Floor -
Reciprocal of Session
_Leneth

10 20 30 40 50 60
CONSECUTIVE SESSIONS
Figure 4. Baseline, extinction, and experimental phase where right
leg-movement reinforced: Experiment II


Baseline


.01

.005



.001


_I I I





1000

500



100

50


.01

.005



.001


10 20 30 40 50 60
CONSECUTIVE SESSIONS
Figure 5. Baseline, extinction, and experimental phase where left
leg-movement reinforced: Experiment II






TABLE 3


Adjusted and Unadjusted Mean Rates forEach Phase: Experiment II


Leg-Movement Baseline Experimental Extinction
(Unadjusted) D (Unadjusted) (Unadjusted)
Red (S Left) Green (SD Right)


Right 4.72 4.61 4. 70 4. 52

Left 7.43"** 7.95 6.22 4.40


Leg-Movement Baseline Experimental Extinction
(Adjusted) (Adjusted) (Adjusted)
Red (SD Left) Green (SD Right)

Right 0.00 -0.11 -0. 02** -0.20


Left 0.00 0. 52* -1.21 -3. 03


*p 30.
:*p 10.
***p- .01.









to an experimental mean of 7. 95. In the same condition, right leg-

movement rates decreased slightly from a baseline mean of 4. 72

to an experimental mean of 4. 61. During this condition, left leg-

movement exceeded right leg-movement as expected and there were

only two sessions where the right leg-movement rate exceeded the

left leg-movement rate.

In the green light experimental condition, the unadjusted means

do not differ in the expected directions. Although right leg-movement

was reinforced, the left leg-movement mean rate remained higher than

the right leg-movement mean rate. Right leg-movement decreased

slightly from its baseline to an experimental mean of 4. 70 during

this component. Left leg-movement also decreased but from its

baseline mean to an experimental mean of 6. 22.

Analysis of the data shows an initial reliable difference between

the baseline means with left leg-movement exceeding right leg-move-

ment (t=3. 14, pZ. 01). Because of this initial difference, adjustment

of the means causes a reversal of the means in the green light experi-

mental component so that they now fall in the expected direction. With

adjustment, all experimental means fall in the expected directions.

This indicates movement toward, but not establishment of, discrim-

inative responding. In the presence of the red light, leftleg-movement

exceeds right leg-movement but the differences between the means are

not significant (t_=1. 09, p_. 30). In the presence of the green light,








adjusted right leg-movement exceeds adjusted left leg-movement

but again the differences between the means are not significant

(t=1.78, p. L .10).


Experiment III

Figure 6 illustrates the results from this differentiation study.

Only one graph is used because there is never more than one condi-

tion occurring in any one phase of the differentiation studies. Table 4

lists the adjusted and unadjusted means for each phase of this

experiment.

Observation of the graph shows that right leg-movement remained

at a higher rate than left leg-movement during the eight session base-

line. There was only one session where the left leg-movement rate

exceeded the right leg-movement rate. During the thirteen sessions of

experimental phase I, this pattern remained the same although left

leg-movement was being reinforced. In this condition there were

three sessions where left leg-movement exceeded right leg-movement.

The first extinction phase of four sessions shows an initial depression

of rates followed by a one session rise and another session where

rates fell again below baseline.

During the eleven sessions of experimental phase II, right leg-

movements were reinforced and the initial baseline differences were

enhanced in the expected directions. There were no sessions where





1000
500


100
50 Base. Experimental Ext. Experimental Ext.
Sr Left Sr Right

C 10
5

.D 5 5 I


;,: A,1



I
i .1 IA/^7 v *-
.05 .



.01 Right Leg-Movement L
.005 ------ Left Leg-Movement
,_r-- Record Floor -
Reciprocal of Session
.001 .Length
10 20 30 40 50 60
CONSECUTIVE SESSIONS
Figure 6. All baseline, experimental, and extinction phases:
Experiment III







TABLE 4


Adjusted and Unadjusted Mean Rates for Each Phase: Experiment III


Leg- Baseline Experimental I Extinction I Experimental II Extinction II
Movement (Unadjusted) (Sr Left) (Unadjusted) (Sr Right) (Unadjusted)
(Unadjusted) (Unadjusted)

Right 2.39** 2. 15 0.92 2.41 2.04

Left 1. 54 1. 53 0. 96 0. 78 1. 33



Leg- Baseline Experimental I Extinction I Experimental II Extinction II
Movement (Adjusted) (Sr Left) (Adjusted) (Sr Right) (Adjusted)
(Adjusted) (Adjusted)

Right 0.00 -0.25 -1.48 0.35*** -0.36


Left 0.00 -0.01* -0.59 -0.85 -0.21


p .. 50.
**p" 10.
**"*p 01.





52
*


left leg-movement rate exceeded right leg-movement rate. The

four sessions of final extinction data show a similar pattern to

that illustrated in the first extinction phase without the final drop

in rates.

Experimental phase I shows the mean unadjusted rates in reverse

to the expected directions. Left leg-movement was being reinforced

but its mean rate remained about the same as its baseline mean and

below that of right leg-movements. Left leg-movement rates went

from a baseline mean of 1. 54 to an experimental mean of 1. 53.

Right leg-movement rates decreased slightly in this condition from

a baseline mean of 2. 39 to an experimental mean of 2. 15.

After four sessions of extinction where left leg-movements

decreased to a mean rate of .96 and right leg-movements to a mean

rate of .92, the second experimental phase began. Right leg-move-

ments were reinforced during this phase and increased from a base-

line mean of 2. 39 to an experimental mean of 2. 41. Left leg-move-

ment mean rates decreased from their baseline mean of 1. 54 to an

experimental mean in this condition of .78. Final extinction showed

left leg-movement rates increasing back to approximately their level

of baseline and right leg-movement rates decreasing somewhat to a

level below their baseline.

Adjustment of the experimental means again leaves both sets of

the experimental data in the expected directions. During experimental









phase I when left leg-movements were reinforced, the unadjusted

means show no conditioning but the adjusted means occur in the

expected direction to suggest a movement toward that conditioning.

Adjusted left leg-movements occurred slightly more frequently than

adjusted right leg-movements, but the differences between their

means was not significant (t=. 69, pt 5). In experimental phase II,

the means remained in the expected direction after adjustment.

Right leg-movements were reinforced and they occurred more

frequently than left leg-movements. The differences between the

experimental means in this condition were significant (t=3. 33,

pl .01). Initial baseline mean differences approached but did not

reach significance (t=2. 34, p< 10). This analysis showed that

differentiated responding occurred when the leg-movement preference

from baseline was emphasized but not when a reversal of that base-

line preference was attempted.


Experiment IV

Figure 7 illustrates the results from this concluding differen-

tiation study. One graph is used for the same reason as was

explained in the results section for Experiment III. Table 5 lists

the adjusted and the unadjusted means for each phase of this

experiment.




1000

500



100
50


10
5 -I


1

.5 -


.1 -
.05 -


Base.


Exp.
S Right







S ^


Ext.








I

/\ *
F,
'i
/-V


Exp.
- Left


.01
.005


.001


Ext.





i i
S 1


Right Leg-Movement
. Left Leg-Movement
x-y-x Record Floor -
Reciprocal of Session
Length


10 20 30 40 50 60


CONSECUTIVE SESSIONS
Figure 7. All baseline, experimental, and
Experiment IV


extinction phases:


I .


L


~lici






iiiL








TABLE 5


Adjusted and Unadjusted Mean Rates forEach Phase: Experiment IV


Leg- Baseline Experimental I Extinction I Experimental II Extinction II
Movement (Unadjusted) (Sr Right) (Unadjusted) (Sr Left) (Unadjusted)
(Unadjusted) (Unadjusted)

Right 6.51** 5. 76 5.93 5.03 3.34


Left 5.40 3.34 2.80 4.59 4.01


Leg-, Baseline Experimental I Extinction I Experimental II Extinction II
Movement (Adjusted) (Sr Right) (Adjusted) (Sr Left) (Adjusted)
(Adjusted) (Adjusted)

Right 0.00 -0.75*** -0. 58 -1.48 -3.17


Left 0.00 -2.06 -2.60 -0.81* -1.39


*p d .40.
**p 20.
***pds. 01.









The graph illustrates that right leg-movement exceeded

left leg-movement during the five session baseline; there was

only one session where left leg-movement surpassed right leg-

movement. Both left and right leg-movement decreased from

baseline during the ten sessions of experimental phase I where

right leg-movementswere reinforced and remained above left leg-

movement during this phase. There were no sessions where left

leg-movement surpassed right leg-movement. The first extinction

phase of five sessions caused an immediate rise in both rates fol-

lowed by a subsequent drop in responding.

During the eleven sessions of experimental phase II both

left and right leg-movement rates remained approximately equal

to each other, and although left leg-movements were reinforced,

right leg-movements remained somewhat higher in rate across

all sessions. There were five sessions where left leg-movement

exceeded right leg-movement; three out of the last four sessions

contained left leg-movement dominance. Left leg-movement con-

tinued to vary during the five session final extinction phase while

right leg-movement declined steadily.

During experimental phase I when right leg-movement

was reinforced, the unadjusted means show the right

leg-movement rate decreasing from a baseline mean of

6. 51 to an experimental mean of 5. 76. Left leg-movement









also decreased during this condition but from its baseline mean of

5.40 to an experimental mean of 3. 34. These means occur in the

expected direction. After increasing to a mean of 5. 93 during

extinction phase I, right leg-movement rates decreased to a mean

of 5. 03 during the second experimental phase. Left leg-movements

were reinforced during this phase and increased from their extinction

mean of 2. 80 to an experimental mean of 4. 59. The left leg-move-

ment mean rate is below the right leg-movement mean rate during

this phase so the means do not fall in the expected directions.

However, the means are closer during this phase than they were in

any preceding phase. During final extinction, left leg-movement

fell to a mean of 4. 01 and right leg-movement fell to a mean of 3. 34.

After adjustment, both sets of experimental means occur in

the expected direction. During experimental phase I, the adjusted

means occur in the expected direction with right leg-movement

rates significantly higher than left leg-movement rates (t=3.42,

pL .01). The adjustment of the experimental phase II means caused

a reversal in left-right placement so that they, too, occur in the

expected direction. Left leg-movement rates occurred slightly

more than right leg-movement rates but the differences between

the means were not significant (t-1. 04, p/ 40). Initial right over

left baseline differences were not significant (t=1. 75, p..20). As

in Experiment III, emphasis of the initial baseline preference was

successful while reversal of that preference was not.





58 *


Summary

Four experiments were completed investigating complex

learning in the human infant. All experiments studied the effects

of contingent visual stimulation on the leg-movement response.

Discriminative responding developed under both conditions in

Experiment I. No discriminative responding was found in

Experiment II. In Experiments III and IV differentiated responding

was found in only one of the two experimental phases in each case.

In all cases the adjusted experimental means fell in the expected

directions indicating movement toward appropriate differential

responding. These results and their implications are discussed

in the next chapter.















CHAPTER FIVE

DISCUSSION AND CONCLUSIONS


Discussion


The results of these studies on response differentiation and

stimulus discrimination are consistent with, and add credence to,

a conditioning view of learning. Some of the outcomes presented

here deserve special consideration and are thus discussed in this

section. These outcomes include the failure of the reinforcement

used to increase the rates of leg-movement, the accompanying

development of differentiated and discriminative responding with-

out that increase, and the different trends toward conditioning in

the two types of tasks when compared to initial baseline preferences.

This latter outcome may allow for some suggestions to be made

concerning differences between discrimination and differentiation

as forms of learning.

The failure of the reinforcing stimuli in increasing leg-movement

rates is of primary importance in explaining why significant differences

did not occur in some of the experimental phases. The concept of

reinforcement is functionally defined,and where leg-movement rates





60 *


failed to increase, the stimulus consequences did not actually act

as reinforcers. Without this occurrence, it is expected that neither

differentiation nor discrimination would occur. In general, neither

the conjugate mobile reinforcement nor the light flashes served as

very adequate reinforcers.

Experiment I, however, is the exception since the conjugate

mobile reinforcement was effective and the audible discriminative

stimuli came to control responding. Experiment II was identical

to Experiment I except for the substitution of visual discriminative

stimuli for the audible ones. The lack of significant results in

Experiment II may be partially attributed to some form of modal

interference between the visual discriminative stimuli and the

visual reinforcement which may have weakened the effectiveness

of the reinforcer. Casual observation of the infant during this

experiment prompted this speculation because the infant seemed to

spend more time observing the lights than he did the mobile. This

casual observation was also what promoted a change from mobile

reinforcement to light flashes as reinforcement in the last two

experiments. No other research is available to support that con-

tention of modal interference but the observed behavior of the

infant deems that reasoning plausible.

Light flashes as a reinforcement did not seem especially

powerful either. In Experiment III, one light flashed as a









consequence to leg-movement and conditioning occurred in only

one of the experimental phases. Two additional lights were added

to the apparatus in Experiment IV in an attempt to increase the

effectiveness of light flashes as a reinforcer. Again, conditioning

occurred in only one of the experimental phases. Although condi-

tioning occurred, there was only one slight increase from baseline

in any of the experimental phases. That occurred in Experiment III

where only one light was used as the reinforcer. Even though some

differentiated responding occurred, light flashes, as used here,

did not constitute a very effective reinforcer.

Two outcomes of these studies are of special interest in light

of the lack of reinforcer effectiveness. The first outcome concerns

the fact that differentiation and discrimination developed, or moved

toward development, without acceleration of the rates of the leg-

movement being reinforced. This finding holds for all the studies

except Experiment I where all experimental phase rates increased

from baseline. No experimental means in Experiments II, III, and

IV increased appreciably from their baseline means. In two cases

experimental means surpassed baseline means but these differences

were slight. Still, statistical significance was found in one condition

in both Experiments III and IV, and was approached in one condition

in Experiment II. The data demonstrate that,in all of these cases,

differentiated and discriminative responding developed from suppression





62 *


of the leg-movement rate not receiving reinforcement, rather than

from acceleration of the leg-movement rate being reinforced. In

the cases where statistical significance was not achieved, but the

adjusted experimental means fell in the expected directions, the

same finding held. Conditioning occurred, or tended toward

occurrence, in a manner reverse to that which would be expected.

The reinforcement seemed to maintain responding around baseline

levels, while extinction depressed it.

The second outcome of interest, and possibly the most impor-

tant, is the occurrence of different trends toward conditioning in

the two types of studies. In all the studies reported here, there

was always a baseline preference in leg-movement; in all but

Experiment II, the preference was right leg-movement over left

leg-movement. The different types of studies showed different

trends in regards to either reversal of this baseline preference or

enhancement of it. In the discrimination studies, reversal of this

baseline preference was quite apparent. The results of Experiment I

showed a greater difference between adjusted means in the experi-

mental phase when the contingencies demanded a baseline preference

reversal than when they demanded a baseline preference enhancement.

Although no difference between the means were statistically significant

in Experiment II, differences between the adjusted experimental means

were greater when a reversal was demanded than when enhancement









was called for. It was under this condition in Experiment IIthat

statistical significance was approached (p& 10).

The opposite was true for both of the differentiation studies.

In these two studies, the differences between the adjusted experi-

mental means were greater when the contingencies called for a

baseline preference enhancement. Reversal of the baseline

preference was found after adjustment when demanded by the

contingencies but the differences between the adjusted reversed

means were small. This differential responding in terms of

baseline preferences may, with further study, suggest some

reliable differences between discrimination and differentiation

tasks in regard to infant learning.

Stimulus discrimination incorporates response differentiation

and,in that sense, Experiments III and IV can be seen as having

pulled the differentiation component from Experimentsland II.

The experimental phases of the differentiation experiments

where baseline preference was enhanced demonstrate that

differentiated responding is possible in human infants. The

addition of the discriminative stimuli in the discrimination experi-

ments, especially in Experiment I, seems to have helped this

differentiation to occur. In other words, differentiation seems

to be assisted by the discrimination paradigm; the addition of stimulus

control may have made the differentiation task more readily learned.









This may suggest that discrimination is a less complex task

than differentiation. This possibility, along with the findings

concerning baseline preferences and conditioning without rate

increases, requires more research before more definite statements

can be made concerning these aspects of infant learning of response

differentiation and stimulus discrimination.


Conclusions

In general, it seems possible to establish discriminative and

differentiated response repertoires in young infants. The results

of the studies presented here, where statistically significant condi-

tioning occurred in four of the eight experimental conditions and

was approached in one other, support that contention as do results

presented elsewhere (Sheppard, 1969; Caron, 1967; Levison &

Levison, 1967). The conditions necessary to establish those

repertoires include adequate instrumentation, effective reinforce-

ment, and, of course, access to infants as subjects. The latter

condition can be the most difficult to implement. These three

conditions and conclusions drawn from their discussion are

presented in this section.

Instrumentation problems are easily solved. After a decision

has been made concerning which response, or responses, an

experimenter wishes to investigate, designing an effective









transducer for that response is relatively simple. Also, there

are many apparatus articles in the literature discussing equipment

that has proven to be useful in the past (DeLucia, 1967; Simmons &

Lipsitt, 1961; Kron, Stein, & Goddard, 1963; Crowell, Peterson, &

Safely, 1960).

Effective reinforcement remains a problem. The visual rein-

forcers used in the studies presented here seemed to lack adequate

strength. Other studies, too, have used visual reinforcement and

ended with negative, or partially negative,results (Friedlander, 1961;

Watson, 1969; Lipsitt, 1963). Visual reinforcement can be effective,

however, and was effective under the circumstances used in

Experiment I. Many studies have used visual reinforcement with

success (Rheingold, Stanley, & Cooley, 1962; Rovee & Rovee, 1969;

Watson, 1970; Lipsitt, Pederson, & DeLucia, 1966). It seems,

however, that for visual reinforcement to be most effective, it must

consist of a wide variety of stimuli which are often interchanged

(Koch, 1967).

From the review of the literature, it seems that the most

promising type of reinforcement for future studies would be

consumable reinforcement. The only problem which could arise

with consumable reinforcement concerns the area of food deprivation.

Food deprivation is both medically and morally unsound when used with

young infants, but studies have been reported without resorting to





66 *


its use. No study found, in fact, reported insignificant results

when consumable reinforcers were employed, and food deprivation

was never used. In light of this evidence, it seems that the use

of consumable reinforcers holds the most promise for studies in

the future.

The use of operant conditioning techniques with children other

than the experimenter's own raises the problem of access. This

writer interviewed at least three sets of parents for every one

who agreed to participate. This is a problem, it seems, for two

reasons. First, the unusually long time length of most operant

research necessitates hardships on parents not extremely interested

in the results of the study. Second, and more difficult to overcome,

some parents are worried that such research will in some way

injure their child. The use of electrical apparatuses is a prime

cause of that concern. This problem can be overcome, however,

and relative to other problems becomes one of minor annoyance

to the experimenter.

But access is only one of the problems encountered when work-

ing with young human infants. Infants have a tendency to do many

things which interfere with the efficient execution of the experiment.

For one thing, infants cry; this becomes a problem when working

with most responses, but especially so with the leg-movement

response. Anyonewho has casually observed infants can determine









that rapid leg-movement is highly correlated with crying. Increases

in rate for this type of reason can hardly be attributed to reinforce-

ment. Infants also become ill, fall asleep, dampen their diapers,

and commit various other assaults on an experimenter's desire to

rapidly complete his work. But the benefits far outweigh the

detriments for infants also smile, laugh, and supply data far more

suitable to answering questions about human learning than do the

young of any other species.

These studies have been completed to add to the knowledge on

human learning and behavior. Research on infant learning is a

comparatively new area and it is too early to speculate on its

contributions. Still, there are sure to be some potential benefits

which can be derived from further work along the lines of those

presented in these studies. Some of the more promising speculations

follow.

Watson (1966) has speculated that early deprivation of contingency

contacts in the first three months of an infant's life may well cause

intellectual deficits. He mentions that "this natural deprivation may

well produce debilitating intellectual effects by establishment of

inappropriate habits of contingency analysis" (p. 57). Experiences

such as those incorporated into these studies could help overcome

that "natural deprivation. "





68 *


Swanson and Benton (1955) have noted that "for the most part,

psychologists have viewed right-left discrimination as an index of

general intellectual development" (p. 123). They go on to say,

"the interest of neurologists in right-left discrimination is derived

from the observation that some patients with cerebral pathology

manifest a striking impairment of this ability" (p. 123). Similar

studies could be done along the lines begun by Barrett and Lindsley

(1962) except with young infants. These future studies, building on

the left-right discrimination and differentiation studies presented

here, could possibly lead to early identification, prevention, and

reversal of retarded intellectual development and some forms of

cerebral pathology.

Another question for future research was raised by the dif-

ferent results obtained in these studies concerning baseline prefer-

ence reversals. If a reversal is more probable when the contin-

gencies demand discrimination than when they demand differentiation,

one positive difference in the learning of these tasks by infants has

been uncovered. If so, this finding must be taken into account when

future studies along these lines are attempted. There is also the

possibility that these differences in respect to baseline preference

reversals are simply artifacts of these studies and will not be

found when larger samples are used.





69



Beyond potential medical or social benefits, the results

presented here add to the general knowledge of human learning

and behavior. Lipsitt (1963) stated that "any complete theory of

human behavior, and most particularly any theory of human deve-

lopment, should contain information about the learning processes

of the young child" (p. 149). While only half of the experimental

phases in these studies demonstrated conditioning, the data pre-

sented here raised a sufficient number of questions to insure that

these results help build that theory.





































APPENDIX








Form 1: Used to collect baseline and extinction data for
Experiments I and II.


Right Leg Left Leg
Sessions Frequency Rate Frequency Rate Time


1




2


3


4


5


6


7


8


9


10





72*



Form 2: Used to collect experimental data for
Experiments I and II.


Right Leg Left Leg
Sessions SD Frequency Rate Frequency Rate Time


1







3


4


5


6


7


8


9


10





73



Form 3: Used to collect all data for
Experiments III and IV-


Right Leg Left Leg
Sessions SP C Frequency Rate C Frequency Rate Time


1




2


3


4


5


6


7


8



9


10












REFERENCES


Barrett, B. H., & Lindsley, O. R. Deficits in acquisition of operant
discrimination and differentiation shown by institutionalized
retarded children. American Journal of Mental Deficiency,
1962, 67, 424-436.

Bijou, S. W., & Baer, D. M. Child development. Vol. 2. Universal
stage of infancy. New York: Appleton-Century-Crofts, 1965.

Bijou, S. W., & Baer, D. M. Operant methods in child behavior
and development. In W. Honig (Ed. ), Operant behavior: Areas
of research and application. New York: Appleton-Century-Crofts,
1966. Pp. 718-789.

Bijou, S. W., & Sturges, P. T. Positive reinforcers for studies with
children: Consumables and manipulables. Child Development,
1959, 30, 151-170.

Brackbill, Y. Extinction of the smiling response in infants as a
function of reinforcement schedule. Child Development, 1958,
29, 115-124.

Brossard, L. M., & Decarie, T. G. Comparative reinforcing
effect of eight stimulations on the smiling response of infants.
Journal of Child Psychology and Psychiatry, 1968, 9, 51-59.

Caron, R. F. Visual reinforcement of head-turning in young infants.
Journal of Experimental Child Psychology, 1967, 5, 489-511.

Charlesworth, W. R. Persistence of orienting and attending behavior
in infants as a function of stimulus-locus uncertainty. Child
Development, 1966, 37, 473-491.

Crowell, D. H., Peterson, J., & Safely, M. A. An apparatus
for infant conditioning research. Child Development, 1960,
31, 47-51.










DeLucia, C. A. A system for response measurement and reinforce-
ment delivery for infant sucking-behavior research. Journal of
Experimental Child Psychology, 1967, 5, 518-521.

Edwards, A. L. Statistical analysis. New York: Holt, Rinehart,
and Winston, 1969.

Friedlander, B. Z. Automated measurement of differential
operant performance. Paper presented at the meeting of
the American Psychological Association, New York,
September 1961.

Hill, S. D. Performance of young children on three discrimination-
learning tasks. Child Development, 1965, 36, 425-435.

Hulsebus, R. C. The use of audio and visual reinforcers in the
establishment of multiple schedule control of infant operant
behavior. Unpublished manuscript, University of Florida,
1970.

Koch, J. Conditioned orienting reactions in two-month-old infants.
British Journal of Psychology, 1967, 58, 105-110.

Kron, R., Stein, M., & Goddard, K. E. A method of measuring
sucking behavior of newborn infants. Psychomatic Medicine,
1963, 25, 181-191.

Leuba, C., & Friedlander, B. Z. Effects of controlled audio-
visual reinforcement on infants' manipulative play in the home.
Journal of Experimental Child Psychology, 1968, 6, 87-99.

Levison, C. A., & Levison, P. K. Operant conditioning of head
turning for visual reinforcement in three-month infants.
Psychonomic Science, 1967, 8, 529-530.

Lindsley, O. R. Operant behavior during sleep: A measure of
depth of sleep. Science, 1957, 126, 1290-1291.

Lindsley, O. R. Conjugate reinforcement. Paper presented at the
meeting of the American Psychological Association, New York,
September 1961.

Lindsley, O. R. A behavioral measure of television viewing.
Journal of Advertising Research, 1962, 2, 2-12.





76 *


Lindsley, O. R., Hobika, J. H., & Etsten, B. E. Operant behavior
during anesthesia recovery: A continuous and objective method.
Anesthesiology, 1961, 22, 937-946.

Ling, B. C. Form discrimination as a learning cue in infants.
Comparative Psychology Monographs, 1941, 17, 66.

//Lipsitt, L. P. Learning in the first year of life. In L. P. Lipsitt
& C. C. Spiker (Eds.), Advances in child development and
behavior. Vol. 1. New York: Academic Press, 1963. Pp 147-195.

Lipsitt, L. P. Learning processes of human newborns. Merrill-
Palmer Quarterly, 1966, 12, 45-71.

Lipsitt, L. P. Learning in the human infant. In H. W. Stevenson,
E. H. Hess, & H. L. Rheingold (Eds.), Early behavior: Compa-
rative and developmental approaches. New York: John Wiley and
Sons, 1967. Pp. 225-247.

SLipsitt, L. P., Kaye, H., & Bosack, T. N. Enhancement of neonatal
sucking through reinforcement. Journal of Experimental Child
Psychology, 1966, 4, 163-168.

Lipsitt, L. P., Pederson, L. J., & DeLucia, C.A. Conjugate
reinforcement of operant responding in infants. Psychonomic
Science, 1966, 4, 67-68.

Papousek, I-. Experimental studies of appetitional behavior in human
newborns and infants. In H. W. Stevenson, E. H. Hess, & H. L.
Rheingold (Eds.), Early behavior: Comparative and developmental
approaches. New York: John Wiley and Sons, 1967. Pp. 249-277.

Rey, A., & Rey-Pinto, T. Visual discrimination learning in a 10-
month-old child and in dogs. Archives de Psychologic, 1959, 37,
101-125.

Rheingold, H. L., Gewirtz, J. L., & Ross, H. W. Social conditioning v'
of vocalizations in the infant. Journal of Comparative and Physio-
logical Psychology, 1959, 52, 68-73.

Rheingold, H. L., Stanley, W. C., & Cooley, J. A. Method for
studying exploratory behavior in infants. Science, 1962, 136,
1054-1055.









Routh, D. K. Conditioning of vocal response differentiation in v
infants. Developmental Psychology, 1969, 1, 219-226.

Rovee, C. K., & Rovee, D. T. Conjugate reinforcement of infant
exploratory behavior. Journal of Experimental Child Psychology,
1969, 8, 33-39.

Schaefer, H. H. Vibration as reinforcer for infant children.
Journal of the Experimental Analysis of Behavior, 1960, 3, 160.

Sheppard, W. C. Operant control of infant vocal and motor behavior. ---
Journal of ExperimentalChild Psychology, 1969, 7, 36-51.

Simmons, M. W. Operant discrimination learning in human infants.
Child Development, 1964, 35, 737-748.

Simmons, M. W. & Lipsitt, L. P. An operant-discrimination apparatus
for infants. Journal of the Experimental Analysis of Behavior,
1961, 4, 233-235.

Siqueland, E. R. Operant conditioning of head turning in four-month
infants. Psychonomic Science, 1964, 1, 223-224.

Siqueland, E. R., & Lipsitt, L. P. Conditioned head-turning in
human newborns. Journal of Experimental Child Psychology,
1966, 3, 356-376.

Smith, J., Johnson, T.S., & Lindsley, O. R. Project Number
001020069. Kansas City: Behavior Bank, 1967.

Swanson, R., & Benton, A. L. Some aspects of the genetic develop-
ment of right-left discrimination. Child Development, 1955, 26,
123-133.

Todd, G.A., & Palmer, B. Social reinforcement of infant babbling. v
Child Development, 1968, 39, 591-596.

Valentine, C. W. The color perception and color preference of an
infant during its fourth and eighth months. British Journal of
Psychology, 1914, 6, 363-386.

Watson, J. S. The development and generalization of "contingency
awareness" in early infancy: Some hypotheses. Merrill-Palmer
Quarterly, 1966, 12, 123-135.





78 *



Watson, J. S. Memory and "contingency analysis" in infant learning.
Merrill-Palmer Quarterly, 1967, 13, 55-76.

Watson, J. S. Operant conditioning of visual fixation in infants under
visual and auditory reinforcement. Developmental Psychology,
1969, 1, 508-516.

Watson, J. S. Smiling, cooing, and "the game." Paper presented
at the meeting of the American Psychological Association, New
York, September 1970.

Weisburg, P. Social and nonsocial conditioning of infant vocalizations. V/
Child Development, 1963, 34, 377-388.













BIOGRAPHICAL SKETCH


Samuel Michael Deitz was born January 19, 1945, at New

York City, New York. That year he moved to West Palm Beach,

Florida. In June, 1962, he was graduated from Palm Beach High

School. In December, 1966, he received the degree of Bachelor of

Arts in Education from the University of Florida. From 1966 until

1968 he was employed as an instructor by Nova High School in

Fort Lauderdale, Florida. In 1968 he enrolled in the Graduate

School of the University of Florida. He held an NDEA IV Fellow-

ship from 1968 until 1970. In June, 1970, he received the degree

of Master of Arts in Education. From June, 1969, until the present

he has pursued his work toward the degree of Doctor of Philosophy.

Samuel Michael Deitz is a member of Phi Delta Kappa, Kappa

Delta Pi, and the American Educational Research Association.












I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




John M. Newell, Chairman
Associate Professor of Education


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




William B. Ware
Assistant Professor of Education


I certify that Ihave read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




R. E,-/JJster
Associate Professor of Education
/


I certify that Ihave read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




H. S. Pennypacqer
Professor of Psychology












This dissertation was submitted to the Dean of the College of
Education and to the Graduate Council, and was accepted as
partial fulfillment of the requirements for the Degree of Doctor
of Philosophy.

August, 1971





Dean, College f E Zucation





Dean, Graduate School




Full Text

PAGE 1

Operant Behavior in the Hixman Lifant: Differentiation and Discrimination By SAMUEL MICHAEL DEITZ A Dissertation Presented to the Graduate Council of the University of Florida In Partial Fulfillnient of the Requirements for the Degree of Doctor of Philosophy UNIVERSITY OF FLORIDA 1971

PAGE 2

UNIVERSITY OF FLORIDA 3 1262 08552 3578

PAGE 3

This dissertation is dedicated to Byron, Brady, Shannon, and Richard

PAGE 4

ACKNOVfLEDGEMENTS There are many people this writer needs to thank for giving of their time and effort toward the successful completion of a dissertation. Ho"*vever, to list and publicly thank them all is impossible. A few deserve special consideration and this writer would like to acknov/ledge them here. Special appreciation is due to this v/riter's advisory coinmittee. Dr. Jolm M. Nev/ell, as chairman, contributed many hours and much advice wliich is highly valued. Dr. R. E. Jester, Dr. Williain. B. Ware, and Dr. H. S. PennyjDacker also contributed to the completion of this dissertation. All a.re appreciated for their advice and, inost of all, their friendship. The parents of the infants studied in these experiments were of great help. Dr. and Mrs. R. E. Jester, Dr. and Mrs. Barry Guinagh, Mr. a.nd Mrs. William Graves, and Mr. and Mrs. John Todd all extended much help in gathering the data for these studies. Finally, the v/riter extends Iiis gratitude to his parents and sisters for their support and faith in his ability.

PAGE 5

TABLE OF (Contents Page ACKNOWLEDGEMENTS LIST OF TABLES LIST OF FIGURES ABSTRACT CHAPTER ONE INTRODUCTION TWO REVIEW OF THE LITERATURE . . . . Introdiiction Consumable Reinforcers Social Reinf orcers Tactua l Reinforcers Auditory and Visual Reinforcers Sumniary THREE DESIGN AND INSTRUMENTATION ., Introduction Apparatu s Data C ollection Proc e dures Exp erime nt I Exp eriinent II Experiment III Ex periinent IV Summary 9 12 16 17 23 25 25 26 29 31 32 33 34 35 36

PAGE 6

CHAPTER Page FOUR RESULTS Litrod^ic tion Experinient I . Exper iment II Experiment III Experiment IV SumiTta ry FIVE DISCUSSION AND CONCLUSIONS .. Discussion Conclusions APPENDIX Forms used for data collection REFERENCES BIOGRAPHICAL SKETCH 38 38 39 44 49 53 58 59 59 64 70 70 74 79

PAGE 7

LIST OF. TABLES TABLE Page 1 Procedures Summary for Experiments I Through IV 37 2 Adjusted and Unadjusted Mean Rates for Each Phase: Experiment I 42 3 Adjusted and Unadjusted Mean Rates for Each Phase: Experiment II 47 4 Adjusted and Unadjusted Mean Rates for Each Phase: Experiment III 51 5 Adjusted and Unadjusted Mean Rates for Each Phase: Experinient IV 55

PAGE 8

LIST OF FIGURES / Page Figure ' 1. Infant leg -moveiTient apparatus 27 2. Baseline and experimental phase where right leg-iTioven:ient reinforced: Experiment I 40 3. Baseline and experimental phase where left leg-niovement reinforced: Experiment I 41 4. Baseline, extinction, and experimental phase where right leg-inovement reinforced: ExperiiTLent II 45 5. Baseline, extinction, and experimental phase where left leg-movenient reinforced: Experiment II 46 6. All baseline, experimental, and extinction phases: Experiment III 50 7. All baseline, experimental, and extinction phases: Experiment IV 54

PAGE 9

Abstract of Dissertation Presented to the Graduate Council of the University of Florida In Partial Fulfillment of the RequireiTients for the Degree of Doctor of Philosophy OPERANT BEHAVIOR IN THE HUMAN INFANT: DIFFERENTIATION AND DISCRIMINATION By Samuel Michael Deitz August, 1971 Chairman: Dr. John M. Newell Major Department: Foundations of Education Four separate experiiTients were completed examining operant response differentiation and stimulus discrimination in human infants. The purpose of the st\.idies was to determine whether or not infants could learn these tasks and under what circumstances were they most easily learned. Each experiment used one infant between the ages of six and fourteen weeks of age as its subject, the same basic apparatus, and visual stimuli as reinforcers. The response studied v/as left and right leg-iTiovennents which v/ere defined as tense-extend inotions of either leg which covered a distance of at least one inch. Left and right leg-inovements were measured independently by the apparatus employed in these studies. This apparatus also controlled the presentation or removal of the visual reinforcers and the discriminative stimuli. Visual reinforcement consisted of either light flashes or movement of a reflecting

PAGE 10

mobile on a conjugate reinforcement schedule; the discriminative stimuli were both visual and audible. Baseline, experimental, and extinction phases occurred in all studies except Experiment I where final extinction data were unobtainable. Tlie experiinental phases differed in each study. Experiment I examined differentiated responding under a stimulus discrimination paradigm in one inale infant using conjugate mobile reinforcement and audible discriminative stimuli. Experiment II was identical to Experiment I except visual discriminative stiinuli were substituted for audible stimuli. Experiment III examined response differentiation in one female infant using light flashes as the reinforcement. A conditioning, extinction, reconditioning design was used in this experiment. Experiment IV was identical to Experiment III except that one male infant was used and the light flash reinforcement was increased in intensity. The results of these studies showed conditioning developed in half of the eight experimental conditions and was approached in two others. Appropriate discriiriinative responding developed under both conditions in Experiment I but developed under neither condition in Experiment II. In Experiinent II, however, one condition approached appropriate responding. Differentiated responding developed under only one of the two conditions in both Experiments III and IV.

PAGE 11

It was found that the legmovement preference demonstrated during baseline was an imiportant determinant of appropriate responding in all experiments. Enhancement of that preference effected the only conditioning in the differentiation studies. In the discrimination studies reversal of that preference effected more appropriate responding, hi Experiment 11, it was under this reversal contingency that conditioning was approached. The conclusions drawn from these studies included comiTients on potential social and medical benefits as well as statements concerned with future research on infant learning. It v/as mentioned that one important problem for researchers of infant learning is the discovery of reliable and powerful reinforcers. Visual stiinuli seem to lack this effectiveness and the class of consuinable reinforcers was suggested as the best possibility. An iinportant potential benefit discussed was that contingency experiences such as those incorporated in these steadies inay be of significant aid in building young infants' adaptive behavioral abilities. It was also suggested that investigation into left-right discrimination, in light of medical evidence currently available, may provide for early identification, prevention, and cure of behavioral deficits. Finally, it was concluded that, in addition to these potential benefits, investigations into infant learning can help build a complete theory of human development.

PAGE 12

CHAPTER ONE INTRODUCTION Operant conditioning techniques have been effective in the investigation of early infant learning. By discovering a stable reinforcing stimulus and contingently applying that stimulus as a consequence to certain infant responses, changes in the rates of those responses can be obtained and viewed as early, but distinct, forms of learning. After the effective control of response rates has been established through reinforceinent, more complex forms of learning can be investigated through the differential application of that reinforcement. Forms of coniplex learning of special interest to infant studies have incKided response differentiation and stimulus discrimination. In designing an operant experiment using infants as subjects, two considerations are of primary importance. The first is the selection of the response. For best results, the response must be easily repeatable and occur naturally in the infant's repertoire. Second, it is necessary to use an effective reinforcing stimulus which can be made contingent upon the selected response. Having decided upon the response and the reinforcing stimulus, contingent presentation of

PAGE 13

the reinforcing stimulus should permit investigation into the acquisition of both simple and complex behaviors. "Without an effective reinforcing stimulus, these types of investigations are impossible as it is necessary to establish control over the response through reinforcement in order to shape behavior through the use of differential reinforcement. The investigation of more complex learning througli the use of differential reinforcement demands consideration of both response differentiation and stimulus discrimination. Although botti differentiation and discrimination involve responding, discrimination may be a more complex form of learning as it involves responding to specific discriminative stimuli. Both are foriTis of learning and both involve a change in responding due to contingencies of reinforcement. Response differentiation involves responding onlyThis form of learning can be defined as "making one response instead of another" (Barrett & Lindsley, 1962, p. 424). "Whien one response is emitted instead of another it is because that response made is the one that has been more frequently reinforced. Response differentiation can be thought of as one step more complex than simple acquisition, for acquisition only requires an increase in rate of an undifferentiated response. In the studies reported here, response differentiation was demonstrated by differentially reinforcing either

PAGE 14

the left or the right leg-inoveiTients of young infants and observing appropriate changes in the rates of these leg-movements. Stiinulus discrimination also involves responding Taut discrimination deals primarily with differentiated responding in terms of external, preceding, discriminative stimuli. Because one extra variable has been added to the situation, discrimination incorporates differentiation and can be seen as a more complex form of learning than either simple acquisition or differentiation. Discrimination is defined as "responding to one stimulus instead of another" (Barrett & Lindsley, 1962, p. 424). In this form of learning, if reinforcement is made contingent upon a response in the presence of a specific (discriminative) stimulus, responding will increase in the presence of that stimulus and decrease in its absence. In the studies reported here, stimulus dis crimination vifas demonstrated by reinforcing one leg's movement in the presence of one stimulus and reinforcing the other leg's movement in the presence of another stimulus. Appropriate changes in response rates were again used to signify reliable discriniinative acquisition. Barrett and Lindsley' s (1962) findings concerning operant discrimination and differentiation with retardates have iinplications for research on infant learning. It was demonstrated that deficits in the acquisition of operant discrimination and differentiation may be used to make judgements on the capabilities of retarded ckildren.

PAGE 15

4 Since discrimination and differentiation are two component behaviors of complex tasks, they may be used to clarify the question of behavior deficits. Barrett and Lindsley found different patterns of responding on tasks demanding differentiation and discrimination with children having different types of retardation and suggested that further research along these lines would lead to clearer informiation concerning the capabilities of these children. The studies presented here give preliminary information concerning response differentiation and stimulus discrimination capabilities in young infants. These studies are similar to those of Barrett and Lindsley (1962) in that the same types of information were sought. Further investigations along these lines could clarify many questions concerning infant learning capabilities and possibly lead to early identification of behavioral deficits. All of the studies presented here employed the same response, left and right leg-movement, and all of the studies used soiTie form of visual reinforcement. The tense-extend movement of the leg was chosen as the response for investigation for it is already in the infant's response repertoire; it occurs naturally at above a zero base rate; and it can be easily repeated without fatigue As Sheppard (1969) has comir^ented. The leg kick operant in infants has a great deal to recommend it. The response is not subject to multiple sources of control; infants can perform it

PAGE 16

rapidly with little effort, and no apparent fatigue; the operant level is above zero and the maximum response rate is quite high; and a transducer for the response is extremely simple and easily constructed. Given all these characteristics, it would be an excellent response to utilize in future studies (p. 48). Sheppard's reasoning was helpful in choosing the leg-movement response for these investigations. Visual stimuli were chosen as the reinforcers because they have been shov^n to be effective in the past by a number of studies, and because they are easily adaptable to automatic presentation. Visual stiinuli can also be easily varied in intensity and duration thus making comparative efficacy studies feasible. A more complete discussion of these supporting studies and others is presented in Chapter Two. Two types of visual reinforcement were used in these investigations. In the differentiation studies, light flashes of varying intensity v/ere inade contingent upon the appropriate leg-inovenients. This is a conventional method of reinforcement employing a discontinuous reinforcing event. In the discrimination studies, the less conventional naethod of conjugate reinforcement was employed using the continuous presence of a niobile, where movement of the mobile was contingent upon leg-movement.

PAGE 17

6 . Conjugate reinforcement (Lindsley, 1957, 1961, 1962; Lindsley, Hobika, & Etsten, 1961; Rovee & Rovee, 1969) is a reinforcement system where "the intensity of a continuously available reinforcing stimulus varies directly and inimediately v/ith the rate of response" (Lindsley, 1962, p. 3). In these discrimination studies, movement of the mobile, contingent upon leg-movement, rather than its mere presence served as the reinforcing stimulus. lR.apid leg-movement produced rapid movement of the mobile. Thus, the reinforcing property of this stimulus was found in its intensity, and the intensity varied directly with legmovement. Lipsitt (1963), among others, has recogniz,ed the importance of investigating learning abilities of the young infant and has mentioned the lack of current research in that area. "A review of the available data concerning the learning process of the very young child reveals that comparatively fev/ studies are available, and few of these have systematically explored the various parameters known to affect learning in lower animals or the older person'' (Lipsitt, 1963, p. 148). Some aspects of those parameters wexe investigated in these studies; those -were the parameters dealing with infant learning of response differentiation and stimulus discrimination through reinforcement. It is the intention of the studies presented liere to demonstrate that it is possible for infants to learn complex tasks. Specifically,

PAGE 18

they are presented to give preliminary data on the acquisition of left-right leg-movement discrimination and differentiation by human infants. Evidence is also given showing that visual stimuli can function as reinforcers for establishing these complex behaviors. The purpose of these studies is to establish discriminative and differentiated responding in human infants using visual stimuli as the reinforcement.

PAGE 19

CHAPTER TWO REVIEW OF THE LITERATURE Introduction To further the investigation of infant behavior through the use of operant conditioning techniques, it is necessary to identify stimuli that have been used as positive reinforcers and have been shown to be both powerful and stable. It is also important to consider the types of responses that have been studied with these reinforcers. This review of the literature serves to classify the types of reinforcers that have been used in the various infant studies and to point out the type as v>^ell as the level of complexity of the response involved. This review is limited to only those studies using operant conditioning techniques with infants from birth to the first year of life. Many articles have reviewed studies dealing with operant behavior in the first year of the infant's life (Lipsitt, 1963, 1966, 1967; Bijou & Baer, 1966), but only two have presented a plan for grouping reinforcing stimuli into different categories of a classification system (Bijou & Sturges, 1959; Hulsebus, 1970). The benefits of having a common system of classification of reinforcing stimuli

PAGE 20

are many, but primarily that system allows for easy access to information concerning the effectiveness of the different types of reinforcers. Tlie systems developed by Bijou and Sturges (1959) and Hulsebus (1970) do not perfectly adapt in all categories to classifying information on very young infants, so a i-nodification of their categories is used here. For this review, reinforcers are categorized into the follo'wing classes: consumable reinforcers, social reinforcers, tactual reinforcers, and auditory and -v-isual reinforcers. Within each category, responses are discussed according to the following proposed levels of complexity. First, those studies dealing only -with acquisition are discussed, follow^ed by studies dealing v/ith differentiation and finally studies establishing discriminative repertoires. Consumable Reinforcers A w^ide variety of consuiTiable reinforcers has been used in operant investigations with infants. Honey, jam, saccharine, M & M's, bits of food, sugar, dextrose solution, and milk have all proven to be effective reinforcers for infants. The great majority of the studies, however, employed the last two solutions delivered either by nipple or tube. Lipsitt, Kaye, and Bosack (1966) conditioned the sucking response of infaiits under four days of age. A significant increase over baseline rates v^as shown in sticking a 1/4 inch rubber

PAGE 21

10* tube by using the contingent presentation of a 5 percent dextrose solution for five seconds. Sucking rates increased from baseline during conditioning and decreased during extinction. There was no increase in rate over baseline during the same period for a control group. Siqueland (1964) used milk delivered via a standard nipple to demonstrate differentiation in three-month-old infants. Reliable acquisition and extinction of either right or left head turns by differentially reinforcing the response only when made in the desired directioii was demonstrated. No change was found in either direction for a control group. Other studies have demonstrated some type of discrimination training using consumable reinforcers with infants. Siqueland and Lipsitt (1966) used infants under four days of age to demonstrate discriiTiination learning. A 5 percent dextrose solution was used as the reinforcing stimulus and satiation effects were controlled by keeping the time since feeding constant for all subjects. Reinforcement of the response, ipsilateral head rotation, was accomplished by a two second presentation of the solution via a standard nipple. In the beginning of the experiinent, reinforced head turns were "cued" by a touch to the cheek and a tone to the side indicating reinforcement. The aiithors then proceeded to discrimination training; responses in the presence of one tone were reinforced while responses in the presence of another were not. A reversal was

PAGE 22

11 demonstrated and reliable acquisition shown for all conditions. Papousek (1967) also demonstrated discriminative head turn behavior in infants as young as three months. Left head turns were reinforced with milk in the presence of a bell and right head turns were reinforced in the presence of a buzzer. Acquisition of the original discrimination and a reversal were demonstrated. In other studies the same types of procedures w^ere used. Valentine (1914) established color discrimination in fourand eight-monthold infants using honey and jam as the reinforcers. Ling (1941) investigated forin discriniination in infants as young as six months and achieved reliable results. Saccharine was used as the reinforcer in his studies. Rey and Rey-Pinto (1959) established color discrimination in a ten-month-old infant by reinforcing the correct choice with sugar. Hill (1965) used M & M's and bits of food to establish object discriminations in infants as young as twelve months. Consuiijable reinforcers have proven to be quite effective in establishing operant behavior in infants. Although only one study found (Siqueland & Lipsitt, 1966) formally controlled for satiation effects, the results obtained by the various studies seem to indicate that this is not a problem in infant studies. Complex discriniinative responses as well as simple acquisition have been established using consumable reinforcers and these reinforcers, as a class, seem to be highly effective and enduring. There is much proinise for future studies employing the use of consumable reinforcers.

PAGE 23

Social Reinforcers Most human learning seems to take place under the control of social stimuli. Consequently, many studies of infant behavior have used social stimulation as the reinforcing agent. Different types of human verbal responses, the presence or absence of a human experimenter, bodily contact, and facial expressions have all been used as reinforcers in some of the studies reported on early infancy. As a category, social reinforcement contains innuinerable specific stimuli, but the following studies deinonstrate a beginning toward identifying those stimuli which seei"n to be the most important. Three studies in the literature have built upon each other in an attempt to separate those factors in social reinforcement which are, in fact, reinforcing. Rheingold, Gewirtz, and Ross (1959) used social / reinforcement to condition vocal behavior in three-month-old infants. After two days of baseline, the conditioning phase consisted of an experimenter leaning over the crib, expressionless until the infant produced a discrete vocal sound. The experimenter would then change his expression briefly and return to the expressionless condition until another sound was produced by the infant. This phase was followed by two days of extinction. Reliable acquisition and extinction were deinonstrated and the experimenters concluded that a "social event composed of an everyday complex of acts, performed by an adult who is not the caretaker, can function as a reinforcing stimulus" (p. 72). These results have been questioned since the presence of an immobile adult may serve as a discriniinative stimulus for vocal

PAGE 24

13 behavior in the infant or as a type of "social releaser. " "Weisburg (1963) controlled for this possibility in an experiment which stemmed from the above results. The subjects in this experiment were also three months old and the response definition for vocal behavior was the same. He divided his subjects into the following six groups. Group one had no experimenter present. Group two had a constantly immobile experimenter present with a "blank expression" at all times. Group three received noncontingent social reinf orcennent while group four received noncontingent, nonsocial reinforcenienl consisting of a door chime. Group five received contingent social reinforcement and group six received contingent nonsocial reinforcement. Vocalizations increased over baseline only in the group receiving contingent social reinforcement; thus, this condition is the only one where an actual reinforcer was found to be acting. A functional definition of reinforcement disqualifies the other stimulus consequences being called reinforcers. Weisburg concluded that "the results indicated that the initial presentation of an unresponding human did not serve as a discriiTiinative stimulus (S-^) for vocal behavior" (p. 386). Todd and Palmer (1968) further clarified this issue by again using three-month-old infants and the vocalization response. One of the groups received contingent taperecorded verbal stimulation with an adult present (AP) and the second group received the same stiinulation with no adult present (NAP). By using a sound activated relay system, vocalizations produced the tape-recording for five

PAGE 25

14 « seconds. There was no significant difference between the groups during either baseline or extinction, but the AP group increased significantly higher than the NAP group during conditioning. Still, both groups increased significantly over baseline during conditioning. The results indicate that the "primary influence on infant babbling is the human voice" (p. 595). Further clarification of this issue concerning which factors of a social act are, in fact, reinforcing will come from future research. Two other studies were concerned with the comparative efficacy -of different types of social reinforcers. Koch (1967) used fifty infants at the end of their second inonth and compared the strength of five ''* different social consequences on the head rotation response. The three of his five consequences which iiwolved changing stimuli worked the best and he concluded that the strongest reinforcer is a new and changing stimulus. In comparison to stimulus variety, it was found that the mother's voice and face are comparatively weak reinforcers. Brossard and Decarie (1968) compared eight different social -^ consequences while working with the smiling response of infants between four and five months of age. The following stimuli were used in their experiment and are listed from the most to the least effective: picking up the infant, smiling, rocking, picking up and speaking, touching and speaking, speaking, and touching.

PAGE 26

15 Other studies have dealt with simple acquisition and social reinforcement. Charlesworth (1966) conditioned orienting and attending behavior in fiveto ten-monthold infants using social reinforcement. The social reinforcement consisted of a taperecorded human voice saying "ah . . . peek-a-boo" and the presentation of a terry cloth face. There was significant increase in response rate over the baseline period during conditioning and a return to baseline through extinction. Brackbill (1958) has been the only researcher to report the effects of certain reinforcement schedules on extinction data. The smiling response of threeand one-halfto fourand one-half-monthold infants was recorded; the reinforcing stimuli used social and bodily contact. One group received continuous reinforcement (CRF) during the entire conditioning phase of the experiment. The second group received continuous reinforcement at the beginning of the phase but was soon switched to fixed ratio schedules of values two, three and four. The group receiving the intermittent reinforceinent was found to be more resistant to extinction and Brackbill concluded that "intermittent reinforcenient is superior in maintaining continxious performance of a response during extinction" (p. 124). This result has long been established in the operant literature using animals as subjects but Brackbill's study was the first, and to the present, only, to report these findings with human infants.

PAGE 27

16# All of the studies reported above deal -with simple response patterns. Although no studies using social reinforcement have been reported which establish discriminative behavior, Routh (1969) conditioned differentiated vocalizing using social reinforcement. Thirtyinfants between the ages of two and seven months old were divided into three groups. The reinforceinent for each group consisted of a smile, three "tsk" soi.inds, and a light stroke of the infant's abdomen. One group was reinforced only for consonant sounds, a second group only for vowel sounds, and a third group for any vocalization. All groups increased from baseline, and significant response differentiation occurred in the predicted direction. It was concluded that the resialts v/ere consistent with a conditioning approach to beginning language learning. Social reinforcement presents the intricate problem of deter; . -, mining what aspects of a social act are reinforcing. However, from the above literature, it is evident that this class of reinforcers is a viable one with which to work. J£ social stimulation is to be shown as responsible for much of our adult behavior, more studies need to be done along the lines of that by Routh (1969) where the beginnings of complex repertoires were established and maintained. Tactual Reinforcers Few studies reported in the literature used tactual stimuli as the reinforcing consequences of infant behavior. Schaefer (I960) conditioned button pushing in a seven-month-old infant by applying a

PAGE 28

17 soft, vibrating electric toothbrush to the right palm, of the infant for ten seconds contingent upon each button push. The rate of button pushing increased from baseline during conditioning and decreased during extinction. Sheppard (1969), in a study covered fully under Auditory and Visual Reinforcers, conditioned vocalizations and leg kicking using a similar procedure. In the study by Brossard and Decarie (1968) discussed earlier, touching of the infant was found to be the least effective method of reinforcement among the eight different consequences tried. Touching as defined by those authors, however, was simply the application of the palm of the hand to the infant's abdoinen; no tickling or vibration was involved. The two studies discussed above indicate that tactual reinforcement can be effective if it involves moveinent of the tactual stimulus, such as vibration; it is likely that an immobile tactual stiinulus like that employed by Brossard and Decarie (1968) is an ineffective reinforcer for infant behavior. Auditory and Visual Reinforcers Auditory and visual reinforcers have been used extensively in infant conditioning studies for a number of reasons. Tlieir intensity and duration can be controlled; they are easily presented by automatic equipment; and they have been found to be effective over a wide range of responses. Such varied stimuli as tones, mobiles, lights, paf-fcerns,

PAGE 29

1^ slides, and chimes have all been shown to act as reinforcers for infants under one year of age. These stimuli cannot be thought of as primary reinforcers nor have they been systematically paired with primary reinforcers to insure an acquired reinforcing effect. Their reinforcing ability probably occurs because of the effectiveness of what Bijou and Baer (1965) called "general stimulus change." Simple acquisition has been demonstrated using normal and conjugate reinforcennent in infants under one year of age. Rheingold, Stanley, and Cooley (1962) conditioned touching in four -month-old infants by the contingent presentation of colored geometric forms for one and one-half seconds following each touching response. Rovee and Rovee (1969) demonstrated the effectiveness of conjugate reinforcement of the leg -movement response of ten-week-old infants. An Lshaped mobile was always in the infant's line of vision bvit the infant could move the mobile and produce loud clicks by rapid leg-movement. The experimental group received contingent conjugate reinforcement while the control group received noncontingent stimulation. The leg kick rates of the experimental group tripled over baseline while the rates of the control group showed no significant change. Watson (1970) also used a mobile over the head of his eightweek-old infants. The head press response was employed and each press earned a one second rotation of the mobile. The experimental group showed a significant increase in rate while both control groups.

PAGE 30

19 one receiving noncontingent stimulation and the other viewing a stable mobile, showed no increase. Lipsitt, Pederson, and DeLucia (1966) used a conjugate reinforcement system -where the light illuminating a viewing box varied in intensity according to the rapidity of lever pressing. Lever pressing rates in tv^^elve -monthold infants increased from baseline during both conditioning and reconditioning and decreased during extinction. Smith, Johnson, and Lindsley (1967) conditioned a three-month-old infant to look at another person's eyes by contingently applying the multiple stimulus of a smile and a hand clap on each change in "looks. " Responding increased from baseline during conditioning and decreased again during extinction. Three studies have demonstrated differentiated infant responding using auditory and visual reinforcers. Watson (1967) used both auditory and visual reinforcement to alter the direction of visual fixation in thirteenand fourteen-week-old infants. In two experiments, one with twenty-seven infants and the other with t\venty-four younger infants, a significant increase in differential spatial fixation was obtained during the conditioning phases as compared with the baseline phases. Caron (1967) used a broad range of visual stimuli as reinforcers while working with the head rotation response in threeand one-halfmonth-old infants. An increase over baseline was demonstrated, followed by extinction and reconditioning of the response. Differentiation was demonstrated with some subjects by conditioning a reversal fron:L right head rotation to left head rotation. These results lend statiis

PAGE 31

20 to sensory events as reinforcers and, as she surjainarized, "this research furnished additional evidence that exteroceptive feedback can promote and sustain behavior in infancy" (p. 509). Levison and Levison (1967) also demonstrated differential conditioning of head turning in three-month-old infants. The visual reinforcement consisted of a variety of slides projected on a screen. They conditioned head turning to one side, then after an extinction period reversed the head turning to the other side. Agreeing with Caron (1967), Levison and Levison (1967) conclude that, "rapid conditioning of head turning, extinction and conditioning of a response reversal can be obtained in human infants using visual stimulus patterns as reinforcing stimuli" (p. 530). Three studies have reported failure, or partial failure, in conditioning infant operants using auditory and visual reinforcers. Friedlander (1961) failed to obtain evidence of sLi-Tiple acquisition of cord pulling in two infants, three and seven inonths old respectively, using any of four visual stiinuli as possible reinforcers.. Lipsitt (1963) improved Friedlander 's apparatus in an attempt to condition discriminative behavior in eight-monthold infants but did not demonstrate acquisition of that behaviorWatson (I969) obtained mixed results in his attempt to condition visual fixation in infants from ten to fourteen weelcs of asre. Visual

PAGE 32

21 fixation was differentially reinforced with visual and auditory stinnuli. In the first study, using fourteen-week-old infants, acquisition w^as found for female infants under auditory stimulation but not lander visual stimulation; the opposite was true for male infants. Using ten-week-old infants in a second experiment, consistent learning was found for girls but no significant learning for boys under either, or a coinbination of both, types of stiniulation. The studies involving discrimination learning in infants using auditory and visual reinforcement have been quite successfiil. Leuba and Friedlander (1968) conditioned the discrimination between lighted and unlighted toys in sevento eleven-month-old infants. Wlien the lighted toy was touched, door chiines would sound and varied colored lights would flash. The unlighted toy produced neither auditory nor visual feedback. All twenty infants manipulated the lighted toy more often; it was pressed tvi^ice as often and for three times as long as the unlighted toy. Another discriminative response was established in t-welve-inonthold infants by Si:iimons (1964). Using "tapping" behavior as the response and the sound of door chimes as the reinforcing stimulus, acquisition of a red-blue discrimination was demonstrated. The infants were divided into two groups and placed before a panel with two circles on it; these circles could be illuminated by either a red or a blue light.

PAGE 33

22 * For one group, the red circle signaled reinforcement and the blue signaled extinction. In the other group the program was reversed. The location of the color on the panel, left or right, changed every minute for both groups to control for position effects. Discriminative behavior developed for both groups during conditioning and disappeared during extinction. Sheppard (1969) conditioned two concurrent operants, leg kicking and vocalizations, in an attempt to achieve differential responding under discriminative control in one infant less than three months old. Both operants were placed successfiilly under separate schedule control using the irtother's voice and flashing red and yellow lights as the reinforcing stimuli. At that point, two discriininative stimuli were introduced and responding was conditioned to occur under stimulus control. During this experiment, both differential responding and discriminative responding were achieved, and to the present, this experim.ent is the best demonstration of pure operant research in infants. With just a few exceptions, auditory and visual reinforcers have been shown to be highly effective in controlling infant operant behavior. Because of their flexibility and variety, and because they are easily controlled and presented automatically, auditory and visual stimuli are quite adaptable to operant conditioning experiments. Because of these reasons, reinforcers of this type have been used in the experiments presented in this report.

PAGE 34

23 Summary A reviev/ of operant conditioning studies on infants in their first year of life demonstrates that operant conditioning is a research technique of considerable importance towards understanding learning in the human infant. It further justifies the statement made by Lipsitt (1967) in concluding one of his reviews on learning in the human infant. "Current data on infant learning suggest that a renewed optimism is warranted with respect to the learning potential of infants, as well as to the feasibility of studying that learning potential" (p. 246). It has been shown that a v/ide variety of reinforcing stiinuli is effective in controlling infant operant behavior. The behavior investigated ranges froin simple response acquisition through differentiation and discrimination. Still, there are many areas where more Vv^ork should be done; specifically, questions such as the following need be answered: Wliich aspects of a stimulus complex are, in fact, reinforcing? Wliich types of reinforcers are the most effective? Can developniental levels be determined from investigating the learning of complex tasks? Can early identification of behavioral deficits aid in overcoming those deficits? Livestigation into response differentiation and stimulus discrimination can begin to answer some of the above questions. No work has yet been done vyhich investigates infant response capabilities with these types of learning. The studies presented here begin to analyze.

PAGE 35

24 with infants, the questions dealt with by Barrett and Lindsley (1962) with retarded children. The results presented here are preliminary, but suggest directions in which future research on infant learning could move.

PAGE 36

CHAPTER THREE DESIGN AND INSTRUMENTATION Introduction Four separate experiments, each using visual reinforcement and the same basic apparatus, were conducted to investigate differentiation and discrimination in the human infant. The operant response used for these investigations was the leg-movement response. Legmovement was defined as a tenseextend motion of the leg vi'liich covered a distance in either direction of at least one inch. Each experiment used one infant chosen from those wliose parents volunteered them for these studies. The only criteria for subject selection were that the infants be full term births and that they be judged medically healthy at the time the study began. The ijifants were between six and eight weeks old when the studies began and had not reached the end of their fourteenth week v/hen the studies ended. The unusually long time interval for these studies allowed for experimental adjustinents during the investigations as they were found necessary by inspection of the data. 25

PAGE 37

26 Apparatus The apparatus used for this series of experiments was designed to be adapted to all conditions in the four experiments. This section explains the different possible uses of the apparatus, as well as its dimensions, while the specific methodology for how each experiment was conducted is i-nentioned in a following section. The apparatus (see Figure 1) consisted of a partially enclosed box, twenty-eight inches long, nineteen inches high and twenty-three inches wide. The front, back, and part of the top were open to allow for visual inspection of, and easy access to, the infant. Sides were present to prevent the infant from receiving excess or uncontrolled visual stimulation which could be received by the turning of the head. All interior parts of the apparatus were painted a dull, white color to prevent visual stimulation which might compete with the visual stimuli used for the purposes of the studies. Two boards, each seventeen inches long, projected froin the back of the apparatus and allowed for ribbons to be attached from the ends of the boards, around a fulcrum, to the infant's ankles. Wlien the infant tensed its leg the board was pulled down; springs attached to the board pulled the board back to its original position when the infant extended its leg. The apparatus counted the moveinents of each leg separately in either of two ways. First, mechanical counters were mounted so that when a board was pulled down the lever on the counter was

PAGE 38

27 Figure 1. Infant leg-movement apparatus

PAGE 39

28* depressed and the movement was registered. The leg had to be extended and tensed again in order to have another count registered. Second, microswitches were mounted so that when the infant tensed its leg they were depressed and the count was registered on automatic equipment located on the back of the apparatus. A tense-extend movement of at least one inch in both directions was required to depress either the mechanical counters or the microswitches. Both methods were used in these experiments, though never simultaneously. Due to limitations in the electrical equipment, an inner response time (IRT) of . 75 seconds was demanded for either the count to register or reinforcement to be applied when the microswitch method was being used. This was not found to be a problem, however, due to the somewhat low rate of infant leg -movement. Two types of visual stimuli were made automatically contingent upon leg-movement using this apparatus. First, a mobile was attached to either or both boards controlled by leg -movement to supply conjugate reinforcement. The reinforcing property of the iTiobile was found in the intensity of its inovement rather than merely its presence; the intensity of mobile m.ovement was a direct result of leg-movement. The second type of visual stimulus which was made contingent upon infant leg -movement was a brief flash of colored light. The lights were located underneath the board on which the counters were mounted. Depending on the purposes of the experiinent in progress.

PAGE 40

29 the reinforcing stimuli were made contingent on either left or right' leg-movement, leaving movements of the other leg on an extinction schedule. Tv/o types of discriminative stimuli were also available for use with this apparatus. The first type of discriminative s-fcimulus used was audible stimuli presented through a loudspeaker located behind the infant's head. Audible reinforcers may also have been presented through this loudspeaker but -were not used in these studies. The second type of discriminative stimulus used was visual. These employed red and green colored lights located above the infant's head. These visual stin:iuli were used -when the conjugate naobile reinforcement was used. The flexibility of this apparatus was utilized during the different phases of the four experiments conducted for this study. Soth types of counting i-nechanisms were used as well as both types of reinforcing stimuli, flashing lights for one type of reinforcement and a reflecting mobile for the other. Both audible and visual discriminative stimuli v/ere also used in these studies, a buzzer during, one experiment and red and green lights during another. Dat a Collection Baseline, conditioning, and extinction phases occurxed in all but one experiment where extinction data were not collected. The conditioning phases differed somewhat from experiment to experiment.

PAGE 41

30' Data were obtained in each phase with the infant being placed in the apparatus with the appropriate contingencies in effect for an experimental "session. " Sessions occurred at least once per day at approximately the same times; more than one session often occurred in the same day and sometimes a few days would lapse between sessions. A session did not take place if any of the following conditions were present in the infant at the time of the session: (1) sickness, (2) sleepiness, (3) crying, (4) on drugs or medication, and (5) too "grumpy. " The length of a session depended upon the availability of the infant and its general condition during the session. If the infant began to cry or fall asleep during the session, the session was terminated immediately. No session was considered representative unless it extended for at least three consecutive minutes. The baseline phases extended for at least four sessions as did the extinction phases. The conditioning phases extended for at least ten sessions or longer depending on the data. Data were collected in terms of average rate of leg-movement per session. One or both of the infant's parents were trained to position the infant properly in the apparatus. For each session, the parent placed the infant in the apparatus, turned the counters to zero at the beginning of the session, and kept the length of time of that session by using a stopwatch provided for that purpose. When the session was completed, the parent marked the frequency of movement

PAGE 42

31 of each leg and the elapsed amount of time of the session on special forms provided for that purpose. Rate data were coinputed from the time and frequency data collected by the parents. The forms that were used for data collection are shown in the Appendix. All forms have columns for left and right leg -movement frequencies, left and right leg-movement rates, and the elapsed time of each session. Form one was used to report baseline and extinction data for Experiments I and II; form two was used in the conditioning phases of Experiments I and 11. Form three was adaptable to all phases of Experiments III and IV, The column designating switch position (SP) indicates which conditions were in effect during that phase; the switch is the one located on the electrical equipment panel on the apparatus. The forms were collected from the parents weekly with equipment or design checks made every three or four days. The experimenter was available at all times for any consultation the parents of the infant required. The experimenter controlled the length and duration of any phases and decided on when phase changes occurred. These decisions were made based on examination of the data. Procedures Four experiments v/ere conducted, each with slightly different goals but all investigated response differentiation or stimulus discrimination in the human infant. Experiments I and II were primarily

PAGE 43

32, concerned with stimulus discrimination but contained elements of differentiation. Experiments III and IV dealt only with response differentiation. The following explanations point out the specific procedures used in each experiment. Experiment I Experiment I examined stimulus discrimination and response differentiation in one male infant. The experiment begaxi when the infant was six-weeks-old and ended at his twelfth week. The response was the leg-movement of both legs and the reinforcing stimulus was the conjugate reinforcement of the reflecting mobile. The apparatus was adjusted to count mechanically for this experiment. Baseline and experimental phases were instituted but final extinction data were unattainable because of illness to the infant. The baseline phase consisted of obtaining rate data on the movement of both legs. The mobile was not present for this phase. Baseline data were gathered for a total of fourteen sessions. The leftright leg-movement discrimination was attempted during the experimental phase. Two conditions occurred during each experimental session. These conditions were presented in a random manner over all ten conditioning sessions to control for sequence effects. In one condition, in the presence of a buzzer, the mobile was attached to the board controlled by the left leg. Under this condition, left

PAGE 44

33 leg-movement was reinforced on a continuous conjugate reinforcement schedule. This condition placed right leg-movement on an extinction schedule. In the second condition, a no-buzzer or silent condition, the mobile was attached to the board controlled by the right leg. Here, right leg-movement was on the continuous conjugate reinforcement schedule and left leg-movement was on an extinction schedule. As inentioned, no final extinction data were obtained. Experiment II Experiment II also investigated left-right leg-movement discriminative and differentiated responding in one niale infant. This studybegan at the infant's sixth week and ended at his fourteenth vv'eek. The reflecting inobile served as the conjugate reinforceinent and counting was accomplished mechanically. The same procedures were used as in Experiment I with discrimination training occurring under two conditions w^ithin the same session. In Experiment II, however, the discriminative stiniuli v/ere red and green lights. Twenty-one baseline sessions were coinpleted where neither the mobile nor the discriininative lights -were present. One condition used in this study's experimental phase consisted of the mobile being placed on the board controlled by the infant's right leg in the presence of a green light. The green light signaled continuous conjugate reinforcement for right leg-movement and extinction for left

PAGE 45

3# leg-movement. The second condition consisted of the mobile being placed on the board controlled by the infant's left leg in the presence of a red light. The red light signaled continuous conjugate reinforcement for left leg -movement and extinction for right leg -movement. Again the conditions were randomly alternated to control for a sequence effect. Twenty experimental sessions Avere held follov/ed by six sessions of a final extinction phase. Experiment III Experiment III used one female subject to investigate differentiation in the leg-movement response. This study began at the infant's seventh v/eek and ended at her fourteenth v/eek. The reinforcer was a brief (. 5 second) flash of red light froin one bulb located above the infant's head. The apparatus for this experiment was set up to count legmovement electrically by depression of the microswitches. Baseline, two experimental, and two extinction phases occurred in this experiment. The baseline and extinction phases were identical; leg-movements were counted but no light flash followed any movement. The first experimental phase consisted of thirteen sessions where left leg-movement was reinforced on a continuous schedule of light flash. During all thirteen sessions only left leg-movement v/as reinforced; right leg-movement was thus placed on an extinction schedule. The second experimental phase consisted of eleven sessions where only the

PAGE 46

35 right leg-movement v/as reinforced on a continuous reinforcement schedule for light flash. Left leg-movement during this phase was placed on an extinction schedule. The original baseline phase consisted of eight sessions; the first extinction phase consisted of four sessions; and the final extinction phase consisted of four sessions. No discriminative stimuli were used during this experiment since response differentiation rather than stimulus discrimination was the object of investigation. Experiment IV Experiment IV also investigated response differentiation in one male infant. It was begun at the infant's sixth week and ended at his twelfth week. The reinforcing stimulus was again a brief flash of light, except, for this experiment, three lights flashed simultaneously. One light remained red; the others were amber and yellow. Counting was again accomplished electrically. Baseline, experimental, and extinction phases were used in Experiment IV as they were in Experiinent III. The only change consisted of the right leg-movement being reinforced in the first experimental phase instead of the left leg-iTiovement. Left leg-movement was reinforced during the second experimental phase; this was a reversal of the sequence which occurred in Experiment III. Baseline consisted of five sessions, the first experimental phase of ten sessions,

PAGE 47

m> extinction of five sessions, the second experimental phase of eleven sessions, and final extinction of five sessions. Again no discriininative stimuli were used since none were called for by the purposes of this study. Summary Four studies were completed to investigate differentiated and discriminative responding in the human infant. Tw^o studies dealt primarily -with stimulus discrimination although they contained elements of differentiation, and two with response differentiation. The same apparatus was used for all four studies. Table 1 summarizes these four studies and describes the phases -which occurred in each.

PAGE 48

61 TABLE 1 Procedures Summary for Experiments I Through IV RESPONSE: Tense-extend movements of the right and left legs Experiment I Objective: Discrimination and Differentiation Reinforcement: Conjugate mobile, CRF Phases: 1. Baseline 2. Experimental A. Buzzer Left leg -movements reinforced / Right leg -movements extinguished B. No-Buzzer Right leg-movements reinforced Left leg-movements extinguished Experiment II Objective: Discrimination and Differentiation Reinforcement: Conjugate mobile, CRF Phases: 1. Baseline 2. Experimental A. Red light Left leg-inovements reinforced Right leg-iTLOvements extinguished B. Green light Right leg-movenaents reinforced Left leg-movements extinguished 3_; Extinction ^____ Experiment III Objective: Differentiation Reinforcement: One light flash, CRF Phases: 1. Baseline 2. Experimental I Left leg-movements reinforced Right leg -movements extinguished 3. Extinction I 4. Experimental II Right leg-movements reinforced Left leg-movements extinguished 5. Extinction II Experiment IV Objective: Differentiation Reinforcement: Larger intensity light flash, CRF Phases : 1. Baseline 2. Experimental I Right leg-movements reinforced Left leg-movements extinguished 3. Extinction I 4. Experimental II Left leg -movements reinforced Right leg -movements extinguished 5. Extinction II

PAGE 49

CHAPTER FOUR RESULTS Introduction The four experiments presented in this study were analyzed separately but similar statistical procedures were employed for each. The data from the individual studies are plotted on semi-log graph paper and described in the text. Mean leg-movement rates were computed for the different phases of the experiments from the average rates of leg-movement per session and are presented in tables accomipanying the text. Tlie data were analyzed using the "t"-test for related samples (Edwards, 1969, pp. 135-137) after the experimental sessions were adjusted for initial baseline differences. The adjustments were made by subtracting the mean baseline rate of the leg-movement under consideration from each of its experimental session rates. Adjusted mean experimental rates were then computed and compared by the use of the "t"-test. Significant differences in the expected direction indicated successful conditioning during that phase. 38

PAGE 50

39 Experiment I_ The results of this study are shovv^n in Figures 2 and 3. Figure 2 contains the fourteen sessions of baseline data and the ten sessions of data concerning that part of each experiment3.1 session -when the buzzer was off and right leg-movements were reinforced. Figure 3 illustrates the same baseline data and the ten sessions of data concerning that part of each experimental session when the buzzer was on and left leg-movement received reinforcement. The different conditions in the experimental phase are presented separately to better illustrate how each condition compares to baseline. Table 2 shov/s both the adjusted and the unadjusted means for each phase of this experiment. The graphs illustrate effective conditioning under both conditions. During the no-buzzer component, v/hen right leg-movement resulted in reinforcement, both right and left leg-movement rates increased from baseline, but right leg-movement remained more frequent than left leg-movement. A reversal can be noted by observing Figure 3; left leg-movement rates exceeded right leg-nnovement rates during the buzzer component. Table 2 also denionstrates that discriminative behavior developed under both conditions. During the no-buzzer con-vponent, the mean right leg -movement rate increased from a baseline of 2. 98 to an experimental mean of 9. 09. Although unreinforced in this component.

PAGE 51

40 .1 c 4 -»--^ o " Or-I u u 0) OJ 0) !5X) CO ^ O 'u n W w W > I— I H :d u w o u
PAGE 52

J ^ L 41 J . JL (1)

PAGE 53

42' Oh O _

PAGE 54

43 left leg-movement rates also increased but from a baseline mean of 2. 14 to an experimental mean of 5. 05. Still, left leg-movement rates remained below right leg-movement rates as would be expected. During the buzzer component, the mean left leg-nnovement rate increased from its baseline of 2. 14 to an experimental mean of 9. 95. Right legmovement rates also increased during this component from a baseline of 2. 98 to an experimental mean of 4. 62. As expected, the right legmovement rates remained below the left leg-movement rates in this component. Although the initial baseline rates showed a right over left preference, there were no statistically significant differences between their means (t=l. 36, p4. 20). Also, there were six sessions during baseline where left leg-movement exceeded right leg-movement. Analysis of the adjusted experimental means supports the contention that reliable discriminative responding developed. During the no-buzzer condition, right leg-movement rates were significantly higher than left legmovement rates (;^=3. 28, p/1.01). There was only one session where left leg -movement exceeded right leg -movement in this component. During the buzzer condition, left leg-iTiovement rates were significantly higher than right leg-movement rates (t_=6. 35, p^^i.OOl), and there were no sessions where right leg-moveinent exceeded left leg-movement.

PAGE 55

44 • Experiment II The outcome of this study is presented in Figures 4 and 5. Figure 4 presents the twenty-one session baseline, six session final extinction, and also the data from that part of each experimental session when the green light was on and right leg-movements reinforced. Figure 5 presents the same baseline and extinction data and illustrates twenty sessions of the data concerning that part of each experimental session when the red light was on and left leg-movements were reinforced. Table 3 shows the adjusted and unadjusted means for each phase of this experiment. Observation of the graph shows differences between leg-movement rates during both baseline and the red light experimental component. There is a left leg-movement preference during baseline which is somewhat overcome during the green light experimental component where right leg-movements are reinforced. During the red light experimental condition, that initial baseline difference is expanded. Extinction data are most similar to that of the first four baseline sessions. During baseline, the left leg-movement niean rate exceeded the right leg-movement mean rate; the differences were significant (t = 3. 14, p^.Ol). There were six sessions where right leg-movement surpassed left leg-movement. In the red light experimental condition, left leg-inovement increased slightly from a baseline mean of 7.43

PAGE 56

45 +->

PAGE 57

46 1 <

PAGE 58

47 u ft u «s W o O 0)

PAGE 59

48 to an experiraental mean of 7. 95. In the same condition, right legmovemient rates decreased slightly from a baseline mean of 4. 72 to an experiniental mean of 4. 61. During this condition, left legmovenient exceeded right leg -movement as expected and there were only two sessions where the right leg -movement rate exceeded the left leg-moveinent rate. In the green light experiinental condition, the unadjusted means do not differ in the expected directions. Although right leg-movement was reinforced, the left leg-movement mean rate remained higher than the right leg-movement mean rate. Right leg-movement decreased slightly froin its baseline to an experimental iTiean of 4, 70 during this component. Left leg-movement also decreased but from its baseline mean to an experimental mean of 6. 22. Analysis of the data shows an initial reliable difference between the baseline means with left leg-movement exceeding right leg-movement (t_=3. 14, p^.Ol). Because of this initial difference., adjustment of the means causes a reversal of the means in the greers light experimental component so that they now fall in the expected dixection. "With adjustment, all experimental ineans fall in the expected directions. Tliis indicates movement toward, but not establishment ctf, discriminative responding, hi the presence of the red light, left leg-movement exceeds right leg-movement but the differences between the means are not significant (t_=l. 09, p^. 30). In the presence of the gxeen light.

PAGE 60

49 adjusted right leg-movement exceeds adj\isted left leg-movement but again the differences between the means are not significant (t=l. 78, p. d. . 10). ; Experiment III Figure 6 illustrates the results from this differentiation study. Only one graph is used because there is never more than one condition occurring in any one phase of the differentiation studies. Table 4 lists the adjusted and unadjusted means for each phase of this experiment. Observation of the graph shows that right leg-movement remained at a higher rate than left leg-movement during the eight session baseline. There was only one session where the left leg-moveiTient rate exceeded the right leg-movement rate. During the thirteen sessions of experiinental phase I, this pattern remained the saine although left leg-moveinent was being reinforced. In this condition there were three sessions where left leg-moveinent exceeded right leg-movenient. The first extinction phase of four sessions shows an initial depression of rates followed by a one session rise and another session where rates fell again below baseline. During the eleven sessions of experimental phase II, right legmiovements were reinforced and the initial baseline differences were enhanced in the expected directions. There were no sessions where

PAGE 61

50

PAGE 62

51 X H w o

PAGE 63

52 left leg-movement rate exceeded right leg-movement rate. The four sessions of final extinction data show a siinilar pattern to that illustrated in the first extinction phase without the final drop in rates. Experimental phase I shows the mean unadjusted rates in reverse to the expected directions. Left leg-movement was being reinforced but its mean rate remained about the same as its baseline mean and below that of right leg -movements. Left leg -movement rates went from a baseline mean of 1 . 54 to an experimental mean of 1. 53. Right leg-movenient rates decreased slightly in this condition from a baseline mean of 2. 39 to an experiniental mean of 2. 15. After four sessions of extinction where left leg-movements decreased to a mean rate of . 96 and right leg-movements to a mean rate of . 92, the second experiinental phase began. Right leg-movements were reinforced during this phase and increased from a baseline mean of 2. 39 to an experiinental mean of 2. 41. Left leg-movement mean rates decreased from their baseline mean of 1. 54 to an experimental mean in this condition of . 78. Final extinction showed left leg-movement rates increasing back to approximately their level of baseline and right leg-movement rates decreasing somewhat to a level below their baseline. Adjustment of the experiixiental means again leaves both sets of the experimental data in the expected directions. During experimental

PAGE 64

53 phase I when left leg-movements were reinforced, the unadjusted means show no conditioning but the adjusted nieans occur in the expected direction to sviggest a niovement toward that conditioning. Adjusted left leg-movements occurred slightly more frequently than adjusted right leg-miovements , but the differences between their means was not significant {t~. 69, p^ . 5). In experimental phase II, the means remained in the expected direction after adjustment. Right leg-moveinents were reinforced and they occurred more frequently than left leg-movements. The differences between the experimental nieans in this condition -were significant (t_=3. 33, p<^ . 01). Initial baseline mean differences approached but did not reach significance (t_=2. 34, p<^. 10). This analysis showed that differentiated responding occurred when the leg-movement preference from baseline was emphasized but not when a reversal of that baseline preference was attempted. Experiment IV Figure 7 illustrates the results froni this concluding differentiation study. One graph is used for the same reason as was explained in the results section for Experiment III. Table 5 lists the adjusted and the unadjusted means for each phase of this experiment.

PAGE 65

54 • s:).nuijAj J;3(£; s:;u9ui3Aop«^

PAGE 66

55 u X (6 w o

PAGE 67

56 9 The graph illustrates that right leg-movement exceeded left leg-movement during the five session baseline; there was only one session v^here left leg-movement surpassed right legmovement. Both left and right leg-movement decreased from baseline during the ten sessions of experimental phase I where right leg-movements were reinforced and remained above left legmovement during this phase. There were no sessions where left leg-movement surpassed right leg-movement. The first extinction phase of five sessions caused an immediate rise in both rates followed by a subsequent drop in responding. During the eleven sessions of experimental phase 11 both left and right leg -mio vein ent rates remiained approximately equal to each other, and although left leg -movements were reinforced, right leg-movements remained somewhat higher in rate across all sessions. There were five sessions where left leg-inovement exceeded right leg-moveinent; three out of the last four sessions contained left leg-movement dominance. Left leg-movement continued to vary during the five session final extinction phase while right leg-movement declined steadily. During experiniental phase I when right leg-movement was reinforced, the unadjusted means show the right leg-movement rate decreasing from a baseline mean of 6. 51 to an experimental mean of 5. 76. Left leg-moveinent

PAGE 68

57;also decreased during this condition but from its baseline mean of 5.40 to an experimental mean of 3. 34. These means occur in the expected direction. After increasing to a iTiean of 5. 93 during extinction phase I, right leg -movement rates decreased to a mean of 5.03 during the second experimental phase. Left leg-movements were reinforced during this phase and increased froin their extinction mean of 2. 80 to an experimental mean of 4. 59. The left leg-movement niean rate is below the right leg-iTiovement mean rate during this phase so the means do not fall in the expected directions. However, the iTieans are closer during this phase than they were in any preceding phase. During final extinction, left leg-movement fell to a mean of 4. 01 and right leg-movement fell to a mean of 3. 34. After adjustment, both sets of experimental means occur in the expected direction. During experimental phase I, the adjusted means occur in the expected direction with right leg-movement rates significantly higher than left leg-inovement rates (t = 3.42, p^ .01). Tlie adjustment of the experimental phase II means caused a reversal in left-right placement so that they, too, occur in the expected direction. Left leg-movement rates occurred slightly more than right leg-movement rates but the differences between the means were not significant (t_=l. 04, pA . 40). Liitial right over left baseline differences were not significant (t=1.75, p^ . 20). As in Experiment III, emphasis of the initial baseline preference was successful while reversal of that preference was not.

PAGE 69

58 • Summary Four experiments were completed investigating complex learning in the human infant. All experiments studied the effects of contingent visual stimulation on the leg-moveinent response. Discriininative responding developed under both conditions in Experiment I. No discriminative responding was found in Experiment 11. In Experiments III and IV differentiated responding •was found in only one of the two experimental phases in each case. In all cases the adjusted experimental means fell in the expected directions indicating movement toward appropriate differential responding. These results and their implications are discussed in the next chapter.

PAGE 70

CHAPTER FIVE DISCUSSION AND CONCLUSIONS Discussion The results of these studies on response differentiation and stimulus discrimination are consistent with, and add credence to, a conditioning vie'w of learning. Some of the outcomes presented here deserve special consideration and are thus discussed in this section. These outcomes include the failure of the reinforcement used to increase the rates of leg-movement, the accompanying development of differentiated and discriminative responding v/ithout that increase, and the different trends tov/ard conditioning in the t-wo types of tasks \vhen compared to initial baseline preferences. Tliis latter outcome may allow for soine suggestions to be made concerning differences between discrimination and differentiation as forms of learning. The failure of the reinforcing stimiuli in increasing leg-movement rates is of priniary importance in explaining v/hy significant differences did not occur in some of the experimental phases. The concept of reinforcement is functionally defined, and where leg -movement rates 59

PAGE 71

60 failed to increase, the stimulus consequences did not actually act as reinforcers. Without this occurrence, it is expected that neither differentiation nor discrimination would occur. In general, neither the conjugate mobile reinforcement nor the light flashes served as very adequate reinforcers. Experiment I, hov/ever, is the exception since the conjugate mobile reinforcement -was effective and the audible discriminative stimuli caine to control responding. Experiment II\^'as identical to Experiment I except for the substitution of visual discriminative stiiTLuli for the audible ones. The lack of significant resxalts in ExperiiTient II may be partially attributed to some form of modal interference between the visual discriminative stimuli and the visual reinforcement which may have weakened the effectiveness of the reinforcer. Casual observation of the infant during this experiment prompted this speculation because the infant seemed to spend more time observing the lights than he did the mobile. This casual observation ^vas also what promoted a change from inobile reinforcement to light flashes as reinforcement in the last two experiments. No other research is available to support that contention of modal interference but the observed behavior of the infant deems that reasoning plausible. Light flashes as a reinforceinent did not seem especially powerful either. Li Experiment III, one light flashed as a

PAGE 72

61 consequence to leg-movement and conditioning occurred in onlyone of the experimental phases. Two additional lights were added to the apparatus in Experiment IV in an attenipt to increase the effectiveness of light flashes as a reinforcer. Again, conditioning occurred in only one of the experimental phases. Although conditioning occurred, there was only one slight increase from baseline in any of the experimental phases. That occurred in Experiment III where only one light was used as the reinforcer. Even though some differentiated responding occurred, light flashes, as used here, did not constitute a very effective reinforcer. Two outcomes of these studies are of special interest in light of the lack of reinforcer effectiveness. The first outcome concerns the fact that differentiation and discrimination developed, or moved toward developnient, without acceleration of the rates of the legmoveinent being reinforced. This finding holds for all the studies except Experiment I where all experimental phase rates increased froiTL baseline. No experimental means in Experiments II, III, and IV increased appreciably from their baseline means. In two cases experimental means surpassed baseline means but these differences were slight. Still, statistical significance was found in one condition in both Experiinents III and IV, and was approached in one condition in Experiment II. Tlie data demonstrate that, in all of these cases, differentiated and discriminative responding developed from suppression

PAGE 73

62 « of the leg-movement rate not receiving reinforcement, rather than from acceleration of the leg-movement rate being reinforced. In the cases Vv^here statistical significance was not achieved, but the adjusted experiraental means fell in the expected directions, the same finding held. Conditioning occurred, or tended toward occurrence, in a manner reverse to that which would be expected. The reinforcement seemed to maintain responding around baseline levels, while extinction depressed it. The second outcome of interest, and possibly the most important, is the occurrence of different trends toward conditioning in the t-wo types of studies. In all the studies reported here, there w^as always a baseline preference in leg-inovement; in all but Experiment II, the preference was right leg-movement over left leg-inovement. Tlie different types of studies showed different trends in regards to either reversal of this baseline preference or enhancement of it. In the discrimination studies, reversal of this baseline preference Avas quite apparent. The results of Experiment I showed a greater difference bet\veen adjusted ineans in the experimental phase when the contingencies demanded a baseline preference reversal than when they demanded a baseline preference enhancenient. Although no difference between the means were statistically significant in Experiment 11, differences between the adjusted experimental means were greater \vhen a reversal was demanded than when enhanceinent

PAGE 74

63 was called for. It was under this condition in Experiment II that statistical significance was approached {p^ . 10). The opposite v/as true for both of the differentiation studies. In these two studies, the differences between the adjusted experimental means were greater when the contingencies called for a baseline preference enhancement. Reversal of the baseline preference was found after adjustment when demanded by the contingencies but the differences between the adjusted reversed means were small. This differential responding in terms of baseline preferences may, with further study, suggest some reliable differences between discrimination and differentiation tasks in regard to infant learning. Stimulus discrimination incorporates response differentiation and, in that sense. Experiments III and IV can be seen as having pulled the differentiation component from Experiments! and II. The experimental phases of the differentiation experiments where baseline preference was enhanced demonstrate that differentiated responding is possible in hunian infants. The addition of the discriminative stimuli in the discrimination experiments, especially in Experiment I, seeins to have helped this differentiation to occur. Li other words, differentiation seems to be assisted by the discriinination paradigm; the addition of stimulus control may have made the differentiation task n:iore readily learned.

PAGE 75

This may suggest that discrimination is a less complex task than differentiation. This possibility, along with the findings concerning baseline preferences and conditioning without rate increases, requires more research before more definite statements can be made concerning these aspects of infant learning of response differentiation and stimulus discrimination. Conclusions In general, it seems possible to establish discriminative and differentiated response repertoires in young infantsThe results of the studies presented here, where statistically significant conditioning occurred in four of the eight experimental conditions and was approached in one other, support that contention as do results presented elsewhere (Sheppard, 1969; Caron, 1967; Levison & Levison, 1967). The conditions necessary to establish those repertoires include adequate instrumentation, effective reinforceinent, and, of course, access to infants as subjectsThe latter condition can be the most difficult to impleaTient. These three conditions and conclusions drawn from their discussion are presented in this section. Instrumentation problems are easily solved. After a decision has been made concerning which response, or responses, an experimenter wishes to investigate, designing an effective

PAGE 76

'66 transducer for that response is relatively simple. Also, there are many apparatus articles in the literature discussing equipment that has proven to be useful in the past (DeLucia, 1967: Simmons & Lipsitt, 1961; Kron, Stein, & Goddard, 1963; Crowell, Peterson, & Safely, I960). Effective reinforcement remains a problem. The irisual reinforcers used in the studies presented here seemed to lack adequate strength. Other studies, too, have used visual reinforcement and ended with negative, or partially negative, results (Friedlander, 1961; Watson, 1969; Lipsitt, 1963). Visual reinforcement can be effective, however, and was effective under the circumstances used in Experiment I. Many studies have used visual reinforcement with success (Rlieingold, Stanley, &; Cooley, 1962; Rovee & Rovee, 1969; Watson, 1970; Lipsitt, Pederson, & DeLucia, 1966). It seems, however, that for visual reinforcement to be iTiost effective, it must consist of a wide variety of stimuli which are often interchanged (Koch, 1967). From the review of the literature, it seems that the most promising type of reinforceinent for future studies would be consumable reinforcement. The only problena which could arise with consumable reinforcement concerns the area of food deprivation. Food deprivation is both medically and n-iorally unsound when used with young infants, but studies have been reported without resorting to

PAGE 77

66 its use. No study found, in fact, reported insignificant results when consumable reinforcers were employed, and food deprivation was never used. In light of this evidence, it seems that the use of consumable reinforcers holds the most promise for studies in the future. The use of operant conditioning techniques with children other than the experimenter's own raises the problem of access. This writer interviewed at least three sets of parents for every one who agreed to participate. This is a problem, it seems, for two reasons. First, the unusually long time length of most operant research necessitates hardships on parents not extremely interested in the results of the study. Second, and more difficult to overcome, some parents are worried that such research will in some way injure their child. The use of electrical apparatuses is a prime cause of that concern. This problem can be overcome, however, and relative to other problems becomes one of minor annoyance to the experimenter. But access is only one of the problems encountered when working with young hviman infants. Infants have a tendency to do many things which interfere with the efficient execution of the experiment. For one thing, infants cry; this becomes a problem when working with most responses, but especially so with the leg-movement response. Anyone who has casually observed infants can determine

PAGE 78

67 that rapid leg-movement is highly correlated with crying. Increases in rate for this type of reason can hardly be attributed to reinforcement. Infants also become ill, fall asleep, dampen their diapers, and conimit various other assaults on an experimenter's desire to rapidly complete his v/ork. But the benefits far outvv'eigh the detriments for infants also smile, laugh, and supply data far more suitable to ansv/ering questions about human learning than do the young of any other species. These studies have been completed to add to the knowledge on human learning and behavior. Research on infant learning is a comparatively new area and it is too early to speculate on its contributions. Still, there are sure to be some potential benefits which can be derived from further work along the lines of those presented in these studies. Some of the more proinising speculations follo-w. Watson (1966) has speculated that early deprivation of contingency contacts in the first three months of an infant's life may well cause intellectual deficits. He mentions that "this natural deprivation may well produce debilitating intellectual effects by establishment of inappropriate habits of contingency analysis" (p. 5 7). Experiences such as those incorporated into these studies could help overcome that "natural deprivation. "

PAGE 79

68 Swanson and Benton (1955) have noted that "for the most part, psychologists have viewed right-left discrimination as an index of general intellectual development" (p. 123), They go on to say, "the interest of neurologists in right-left discximination is derived from the observation that some patients with cerebral pathology inanifest a striking impairment of this ability"' (p. 123). Similar studies could be done along the lines begun b^Barrett aoid Lindsley (1962) except with young infants. Tliese future studies, building on the left-right discrimination and differentiation studies presented here, could possibly lead to early identification, prevention, and reversal of retarded intellectual development and some forms of cerebral pathology. Another question for future research was raised by the different results obtained in these studies concexning baseline preference reversals. If a reversal is more probable when the contingencies demand discrimination than when they demand differentiation, one positive difference in the learning of these tasks by infants has been uncovered. If so, this finding must be taken into a,ccount when future studies along these lines are attempted. There is also the possibility that these differences in respect to haseline preference reversals are simply artifacts of these studies and will not be found when larger samples are used.

PAGE 80

Beyond potential medical or social benefits, the results presented here add to the general knowledge of huinan learning and behavior. Lipsitt (1963) stated that "any complete theory of human behavior, and most particularly any theory of hunian development, should contain information about the learning processes of the young child" (p. 149). Wliile only half of the experimental phases in these studies demonstrated conditioning, the data presented here raised a sufficient number of questions to insure that these results help build that theory.

PAGE 81

APPENDIX

PAGE 82

Form 1: Used to collect baseline and extinction data for Experiments I and II. Sessions

PAGE 83

72 Form 2: Used to collect experimental data for Experiments I and 11. Sessions

PAGE 84

73 ForiTL 3: Used to collect all data for Experiments III and IVSessions

PAGE 85

REFERENCES Barrett, B. H. , & Lindsley, O. R. Deficits in acquisition of operant discrimination and differentiation shown by institutionalized retarded children. American Journal of Mental Deficiency , 1962, 67, 424-436. Bijou, S. W. , & Baer, D. M. Child developinent . Vol. 2. Universal stage of infancy . New York: AppletonCenturyCrofts, 1965. Bijou, S. W. , & Baer, D. M. Operant methods in child behavior and development. In W. Honig (Ed. ), Operant behavior : Areas of research and application. New York: Appleton-Century-Crofts, T966. Pp. 7T8^789. Bijou, S. W. , & Sturges, P. T. Positive reinforcers for studies with children: Consumables and manipulables. Child Development , 1959, 30, 151-170. Brackbill, Y. Extinction of the smiling response in infants as a function of reinforcement schedule. Child Developinent , 1958, 29, 115-124. Brossard, L. M. , &; Decarie, T. G. Coinparative reinforcing effect of eight stimulations on the siniling response of infants. Journal of Child Psychology and Psychiatry, 1968, 9, 51-59Caron, R. F. Vis\ial reinforcement of head-turning in young infants. Journal of Experimental Child Psychology , 1967, 5, 489-511. Charlesworth, W. R. Persistence of orienting and attending behavior in infants as a function of stimulus-locus uncertainty. Child Development , 1966, 37, 473-491. Crowell, D. H. , Peterson, J. , & Safely, M. A. An apparatus for infant conditioning research. Child Development , I960, 31, 47-51. 74

PAGE 86

75 DeLucia, C. A. A system for response measurement and reinforcement delivery for infant sucking-behavior research. Journal of Experimental Child Psychology , 1967, 5, 518-521. Edwards, A. L. Statistical analysis . New York: Holt, Rinehart, and Winston, 1969. Friedlander, B. Z. Automated measurement of differential operant performance. Paper presented at the mieeting of the A:nerican Psychological Association, New York, Septeinber 1961. Hill, S. D. Performance of young children on three discriminationlearning tasks. Child Development , 1965, 36, 425-435. Hulsebus, R. C. The use of audio and visual reinforcers in the establishment of multiple schedule control of infant operant behavior. Unpublished inanuscript. University of Florida, 1970. Koch, J. Conditioned orienting reactions in t\vo-month-old infants. British Journal of Psychology , 1967, 58, 105-110. Kron, R. , Stein, M. , 8,; Goddard, K. E. A method of measuring sucking behavior of newborn infants. Psychomatic Medicine , 1963, 25, 181-191. Leuba, C. , & Friedlander, B. Z. Effects of controlled audiovisual reinforcement on infants' manipulative play in the hoine. Journal of Experimental Child Psychology , 1968, 6, 87-99. Levison, C. A. , & Levison, P. K. Operant conditioning of head turning for visual reinforcenient in three-month infants. Psychonomic Science , 1967, 8, 529-530. Lindsley, O. R. Operant behavior during sleep: A measure of depth of sleep. Science , 1957, 126. 1290-1291. Lindsley, O. R. Conjugate reinforcement. Paper presented at the meeting of the American Psychological Association, New York, September 1961. Lindsley, O. R. A behavioral measure of television viewing. Journal of Advertisin g Research , 1962, 2, 2-12.

PAGE 87

76 Lindsley, O. R. , Hobika, J. H. , & Etsten, B. E. Operant behavior during anesthesia recovery: A continuous and objective method. Anesthesiology , 1961, 22, 937-946. Ling, B. C. Form discrimination as a learning cue in infants. Comparative Psychology Monographs , 1941, 17, 66. //Lipsitt, L. P. Learning in the first year of life. In L. P. Lipsitt & C. C. Spike r (Eds. ), Advances in child developmient and behavior. Vol. 1. New York: Academic Press, 1963. Pp 147-195. Lipsitt, L. P. Leariiing processes of human newborns. Merrill Palmer Quarterly , 1966, 12, 45-71. Lipsitt, L. P. Learning in the huinan infant. In H. W. Stevenson, E. H. Hess, & H. L. Rheingold (Eds. ), Early behavior: Comparative and developinental approaches. New York: John Wiley and Sons, 1967. Pp. 225-247. y., Lipsitt, L. P. , Kaye, H. , h. Bosack, T. N. Enhancement of neonatal sucking through reinforcement. Journal of Experimental Child Psychology , 1966, 4, 163-168. J"Lipsitt, L. P., Pederson, L. J., & DeLucia, C.A. Conjugate reinforcement of operant responding in infants. Psychonomic Science, 1966, 4, 67-68. Papousek, H. Experimental studies of appetitional behavior in human newborns and infants. In H. W. Stevenson, E. H. Hess, & H. L. Rheingold (Eds.), Early behavior: Comparative and developmental approaches . New York: John Wiley and Sons, 1967. Pp. 249-277. Rey, A. , & Rey-Pinto, T. Visual discrimination learning in a 10month-old child and in dogs. Archives de Psychologic, 1959, 37, 101-125. Rheingold, H. L. , Gewirtz, J. L. , & Ross, H. W. Social conditioning v^ of vocalizations in the infant. Journal of Comparative and Physio logical Psychology , 1959, 52, 68-73. Rheingold, H. L. , Stanley, W. C. , & Cooley, J. A. Method for studying exploratory behavior in infants. Science , 1962, 136, 1054-1055.

PAGE 88

77 Routh, D. K. Conditioning of vocal response differentiation in i^"^ infants. Developnnental Psychology , 1969, 1, 219-226. Rovee, C. K., & Rovee, D. T. Conjugate reinforceinent of infant exploratory behavior. Journal of Experimental Child Psychology , 1969, 8, 33-39. Schaefer, H. H. Vibration as reinforcer for infant children. Journal of the Experimental Analysis of Behavior , I960, 3, 160. Sheppard, W. C. Operant control of infant vocal and niotor behavior. U'^ Journal of Experimental Child Psychology , 1969, 7, 36-51. Simmons, M. W. Operant discrimination learning in human infants. Child Development , 1964, 35, 737-748. Simmons, M. W. , 8; Lipsitt, L. P. An operant-discrimination apparatus for infants. Journal of the Experimental Analysis of Behavior , 1961, 4, 233-235. Siqueland, E. R. Operant conditioning of head turning in four -month infants. Psychonomic Science, 1964, 1, 223-224. Siqueland, E. R. , & Lipsitt, L. P. Conditioned head-turning in human newborns. Journal of Experimental Child Psychology , 1966, 3, 356-376. Smith, J., Johnson, T. S. , & I.indsley, O. R. Project Number 001020069. Kansas City: Behavior Bank, 1967. Swanson, R. , & Benton, A. L. Some aspects of the genetic developmeait of right-left discrimination. Child Devclop:nent , 1955, 26, 123-133. Todd, G. A. , & Palmer, B. Social reinforcement of infant babbling. ^ Child Development , 1968, 39, 591-596. Valentine, C. W. The color perception and color preference of an infant during its fourth and eighth months. British Journal of Psychology , 1914, 6, 363-386. Watson, J. S. The development and generalization of "contingency av/areness" in early infancy: Sonie hypotheses. Merrill -Palmer Qiiarterly, 1966, 12, 123-135.

PAGE 89

78 • Watson, J. S. Memory and "contingency analysis" in infant learning. Merrill-Palmer Quarterly, 1967, 13, 55-76. Watson, J. S. Operant conditioning of visual fixation in infants under visual and auditory reinforcement. Developmental Psychology, 1969, 1, 508-516. Watson, J. S. Smiling, cooing, and "the game. " Paper presented at the meeting of the American Psychological Association, New York, September 1970. Weisburg, P. Social and nonsocial conditioning of infant vocalization Child Development , 1963, 34, 377-388.

PAGE 90

BIOGRAPHICAL SKETCH Samuel Michael Deitz was born January 19, 1945, at New York City, New York. That year he moved to West Palm Beach, Florida. In June, 1962, he was graduated from Palm Beach High School. In December, 1966, he received the degree of Bachelor of Arts in Education from the University of Florida. FroiTi 1966 until 1968 he was employed as an instructor by Nova High School in Fort Lauderdale, Florida. In 1968 he enrolled in the Graduate School of the University of Florida. He held an NDEA IV Fellowship from 1968 until 1970. In June, 1970, he received the degree of Master of Arts in Education. From June, 1969, until the present he has pursued his work toward the degree of Doctor of Philosophy. Samuel Michael Deitz is a ixieiTLber of Phi Delta Kappa, Kappa Delta Pi, and the American Educational Research Association. 79

PAGE 91

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. vA:,,.^^A \vu..t John M. Ne'well, Chairman Associate Professor of Education I certify that I have read this study and that in niy opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William B. Ware Assistant Professor of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Associate Professor of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertati^on for the degree of Doctor of Philosophy. H. ~S. Pennypaq'Uer Professor of Psychology

PAGE 92

This dissertation was submitted to the Dean of the College of Education and to the Graduate Council, and was accepted as partial fulfillinent of the requirements for the Degree of Doctor of Philosophy. August, 1971 V/(^/%^' Dean, Graduate School