 

FORMULATION AND REF ORMULATION
OF VERBAL CONCEPTS

BY

Eugene Richard Hall

A THE SIS

Submitted to the College of Science and Arts
Michigan State University in partial fulﬁllment
of the requirements for the degree of

MASTER OF ARTS

Department of Psychology

1960

ACKNOWLEDGMENT

The author wishes to eXpress his gratitude and sincere
appreciation for the guidance and assistance in the planning
and execution of this research, and the deve10pment of this
manuscript to Dr. Donald M. Johnson, chairman of his
committee.

In addition, he wishes to convey thanks to Dr. Frank J.
Restle and Dr. Henry C. Smith for their criticism and advice,
during the preparation of this thesis.

*******$***********

ii

FORMULATION AND REFORMULATION
OF VERBAL CONCEPTS

BY

Eugene Richard Hall

AN ABSTRACT

Submitted to the College of Science and Arts
Michigan State University in partial fulfillment
of the requirements for the degree of

MASTER OF ARTS
Department of Psychology

1960

Approved 11;" [1‘7” “ﬁnd/~31,»

ABSTRACT

This thesis was concerned with the generality of a method
of describing and quantifying reformulation behavior in problem
solving. The method consisted essentially of dividing the problem
solving episode into two phases, preparation and solution. During
the preparation phase, the subject saw the part of a problem that
allowed him to form a hypothesis about what the required solution
of the problem would be. He then turned a toggle switch which
illuminated a number of possible solutions. This action prevented
him from seeing the preparation material again unless he turned
the switch back. "Switching-back" was taken as the observable
sign of reformulation.

A previous study using geometric figures about which con-
cepts could be formed, did obtain differences in the switchback
index of reformulation. The present study, designed to assess the
generality of the method, employed verbal concepts. Subjects were
given pretraining on either of two verbal concepts. This pretraining
resulted in the establishment of effective sets but presenting problems
for which the established set was inappropriate did not result in
dispr0portionate returns to the preparation material; consequently,
predicted differences in the switchback index of reformulation did
not appear.

It was suggested that time score increases during the solution

phase could be taken as a sign of reformulation. The increased

iv

time may have allowed the subject to reformulate the problem by
recalling the familiar verbal material of the preparation phase.
If so, this would obviate the necessity of a manual return to

preparation.

INTRODUCTION 0 O O O O O O O O O O O O O O O O O O O O

HistoryofSet. . . . . . . . . .
Current Conceptions of the Nature and Operation

TABLE OF CONTENTS

Of Set I O (O O O O O O I O O O O O O O O O O 0
Advantages Of Set 0 O O O O O O O O O O O I O

Disadvantages of Set
Refo rmulation

EXPERIMENT .

Hypothe si s

De sign . .
Apparatus

Materials

Procedure

RESULTS......

DISCUSSION .

SUMMARY AND CONCLUSIONS .

BIBLIOGRAPHY . . . . . . . . . .

APPENDIX . .

Page

KIO‘U'IUO

11
ll
11
12
13
15
19
28
37
38

41

LIST OF TABLES AND FIGURES
TABLE Page

I Frequency Distribution of Sensory Choices by
Type Of Pretraining O O O O O O O O O O O O O O O O 19

II Errors, Switchbacks and Mean Times in Seconds
OverFiveTestTrials.............. 20

III Time Increase (Log Units) Between Last

Training Trial and First Test Trial . . . . . . . 22
FIGURE
1. Floor plan of apparatuS. . . . . . . . . . . . . . 14

2. Preparation and Solution Time for Training and
T35“: TriaIS--G1‘Oup I. o o o o o o o o o o o o o o 23

3. Preparation and Solution Times for Training and
Test TriaIS--GI‘OUP II o o o o o o o o o o o o o o 24

4. Preparation and Solution Times for Training and
Test Trials--Group III 0 O O O O O O O O C O O O O 25

5. Preparation and Solution Time for Training and
Test TriaIS--G1‘Oup IV 0 o o o o o o o o o o o o o 26

6. Training Trials Performance for 48 Subjects
Having EaCh Concept. 0 O O O O O O O O O O O O O 27

INTRODUCTION

History of Set

 

Near the end of the nineteenth century, extensive investigations
into the nature of thought were carried on at the German University
of Wiirzburg (13, 28, 29). Several important interrelated contributions
towards an understanding of the nature of thinking emerged from this
early psychological laboratory. The Wiirzburg group demonstrated
that factors other than simple associations of conscious ideas or
sensations Operated to inﬂuence the course of problem solving.
These "other" factors appeared to function in such a way that the
thinking processes could be characterized as being "directed" towards
a definite end; namely, the solution of a given problem, and not merely
random processes. In their attempts to account for the direction of
thought the phenomenon of "set" became established (13).

The Wiirzburg group employed a method of study consistent
with the then current concept of thought as the content of conscious-
ness. The "thought experiment" devised by this group consisted of
giving the subject a simple problem to solve and then asking him to
describe his experience from the time when the problem was given
until the solution was reached (28). This introspective method,
however, did not adapt itself well to a systematic analysis of thought.
It was soon discovered that thought could not always be characterized
as having (conscious) content. Certain activities or processes
exhibiting directedness and not Open to introsPective analysis appeared
to be involved in problem solving. The thought processes themselves

leading up to the solution of a problem appeared to be determined

by events other than those which could be reported on by the subject.
One source of this determination was shown to be in the instructions
given by the experimenter (8).

Once a subject had been given a task, he ad0pted a "task-
attitude" or "set" for the duration of the task. This task attitude
determined in large part the resPonses that the subject would make,
the manner in which he would prepare for problem solution and the
nature of the solution that he would give. Ach (as reviewed by
Gibson, 8) early demonstrated this notion of task-attitude by present-
ing the numbers '6 and 2 to his subjects. This presentation yielded a
reaction of 8, 4 or 12 depending on whether the task prescribed was
adding, subtracting or multiplying. The aim, intention or set of
the subject determined the reaction that he would give.

The Wiirzburg investigators expanded this notion of task-
attitude and introduced a terminology employing three principal

concepts. The Einstellung (set) Was produced by the conscious

 

acceptance of the Aufgabe (instructions). The determindierende

 

Tendenz (determining tendency) was a more Specific selective agent

opposable to associative tendencies. The Bewussteinslage (conscious

 

attitude) was the imageless, undescribable experience accompanying
a mental set or trend (8).

Subjective methods of studying problem solving gradually gave
way to objective experimentation. Set now became a hypothetical
construct rather than an intrOSpective observation and the original
precision of meaning for these terms was lost. Many terms now
appear in the literature which all refer broadly to the general fact
that directedness appears as a major characteristic of thinking and

behavior. This directedness may come from Specific inst ructions

as demonstrated by Ach and others (30), or it may arise from the
subject's interpretation of the task properties (23, 18). In this
paper, the term set will be used to refer to directedness in terms

of readiness to make a particular response.

Current Conceptions of the Nature and Operation of Set

 

Problem solving in the laboratory is said to begin with a
stimulus situation and instructions which establish sets and define
the goals (7). Many investigators assume either implicitly or
explicitly that set formation is the major portion of problem solving.
The set steers the course of thinking towards a particular channel,
colors the character of the thought processes and limits the ultimate
possibilities of reSponse (23).

The formation of a set occurs during the period when the
subject is preparing to solve the problem (12, 30). The human per-
ceptual system seen as having a limited capacity (3) requires that a
selective Operation be performed on all inputs into the system. This
selective operation occurs during the preparation period when the
individual assesses, categorizes, codes, groups or otherwise relates
the various parts of the stimulus situation as to their relevancy or
irrelevancy for the solution of the given problem.

The set Operates by selecting in advance. It does not select
from among several responses called up by the stimulus but rather
it limits the field of reSponse in advance of the stimulus so that only
reSponses conforming to the task are ordinarily called up (30). Once
the set has been thoroughly established it may function automatically
with little awareness of its Operation by the subject (28). .

That sets can be established without the use of instructions and

without the awareness of the subjects was demonstrated by Rees and

Israel (23) and later confirIned by Maltzman and Morrisett (l8).
Rees and Israel established sets for certain anagram solutions

over other possible solutions by simply training their subjects on

a particular method of solution. Subjects trained to make solutions
falling into a conceptual category of, for example, nature-words
continued to seek nature-word solutions even when it was possible

to make alternate solutions. Similarly, subjects trained to solve
anagrams by using a certain letter order continued to solve ana-
grams in this manner even when alternative solutions were possible.

The order set was said to be usually involuntary, unverbalized
and unconscious. The category set was found to be either automatic
or accompanied by verbal awareness and a conscious attitude of
search. It was equally effective in either case. The occurrence
of one kind of solution rather than an alternate solution in an
ambiguous situation provided the means of testing the inﬂuence of
the training series in establishing sets toward a certain type of
solution word (category set) or toward a certain method of attack
(order set). The effectiveness of the two kinds of set was measured
in terms of the frequency and quickness of appropriate solutions,

i. e. in terms of the selection and facilitation of solutions.

There seems to be little reason to believe that sets established
by a training method Operate any differently in terms of selectivity
and facilitation from those established by instructions. It may be,
however, that sets established by training are more enduring, more
permanent (19, 26) and Operate at a less conscious level (23) than
the instruction set. The degree of awareness of the set may itself
by a function of the nature of the task.

Set Operates not only in problem solving but also in the related

areas of perception and memory. Kiilpe (15) demonstrated the

effects of set on perception as early as 1904. He found that the
tachistosc0pic presentation of colored nonsense syllables resulted
in perceptions in which the actual sensory qualities of the stimulus
not relevant to the task-set were "to all intents and purposes not
seen" (13, 28). Varying the instructions resulted in different
sensory contents, thus, "the attentive set served to function as a
selective agent. " Various other experiments on the relation of
set to perception have been reviewed by Gibson (8).

Chapman (4) repeated Kiilpe's experiment to determine if the
irrelevant features were really absent from the perception or
whether they were simply forgotten by the time the experimenter
asked the subject to report something he had not been looking for.
Chapman found that both factors were at work. He found that per-
ceptions actually conformed to the instructions but that the process
continued during primary memory with the irrelevant aspects of
the stimulus fading in imagery. Hence, set Operates on both per-
ception and primary memory; it determines what will be remembered

as well as what will be perceived.

Advantages of Set

 

The chief advantages of set lie in its facilitating and selecting
effect on the problem situation (8, ll, 20, 23). The set Operates to
select the inputs into the perceptual system (3). Because of this
selection the variety and kinds of responses that will be called up
into the situation are limited (30). Perception of only relevant
aSpects of the stimulus situation makes for ready categorization of
these aSpects and accurate, rapid solutions of problems can follow.
Set, in providing readily available responses to certain asPects of

the environment, eliminates the need to find new responses to

recurring every day situations and "frees" the mind to let it deal

more effectively with other more complicated tasks (l6).

Disadvantages of Set

 

But the effects of set are not always beneficial to the subject.
Certain dangers are inherent in set. The existing set may not be
applicable to a given problem or it may lead to a solution inferior
to one which could be obtained by using another strategy. Instead
of being master of his habits, the individual may come to be
mastered by them. Luchins has shown the blinding effects of too
persistent a set (16). He demonstrated the effects of set in direct-
ing the method of solution of his now famous water jar problems.
His subjects were trained on one method of solving the problems which
required the manipulation of three jars in such a manner that a
certain Specified amount of water would be left in one of them.

After having been trained to use a longer, more complex method of
solution, his subjects were "blind" to a shorter more direct method
of solving the problems.

Duncker (6) uses the term "functional fixity" to refer to a
disadvantage of set which depends on learning or transfer. If an
object has one customary use or function, it is not easily seen as
suitable for a different function. In an experiment by Adamson (1)
subjects found it difficult to conceive of boxes already serving one
purpose, namely, that of containing material, to be used for another
purpose. Similar experiments utilizing Maier's two-string problems
also point to this phenomenon of functional fixity (2,18).

Interference in transfer exPeriments may be the result of
conﬂicting sets. In certain situations, antagonistic or conﬂicting

sets may result in delaying (5) or altogether preventing an appropriate

reSponse (l6). Rees and Israel (23) demonstrated the effects of
interaction between two kinds of set on both selection of and Speed
of solutions. Using a category set (nature words) and an order-
set, they found that when the two sets were congruent, the combi-
nation increased the number of set solutions. When the two sets
were in opposition, the order set was more effective in determining
solutions. Time scores remained relatively short in the critical
series when the two sets were congruent but there was still some
increase over the time scores of the training series. When the two
sets were in Opposition, time scores for the critical series were
considerably lengthened. Rees and Israel concluded that the
presence of the one set (nature set) tended to hinder the Operation
of the other (letter order set).

Failure in problem solution is seen as being frequently due to
an inﬂexible direction which blocks the correct way of looking at
the problem (8, 17). If set is assumed to be the primary portion of
problem solving, then a correct solution to a problem depends upon
the extent to which the subject has formed an appropriate set
(as well as upon the nature of the problem) in his preparation. If
the subject discovers that he is set wrong for the solution of a
problem he must discard the inappropriate set, return to a new
preparation for the problem and reformulate the problem in terms

leading to a new more apprOpriate set.

Refo rmulation

 

One way of approaching the study of the inappropriate set is
through a study of the reformulation process. What happens when
the person finds that his set is wrong, that he is not prepared to solve
a simple problem? What behaviors of the subject can be used to

infer and describe the reformulation process?

Reformulation can be assessed subjectively by asking the
subject when the experiment is over to comment on his "experience"
at finding that he was not prepared to solve a simple problem. This
method has among its disadvantages one of being temporally
separated from the problem on which reformulation occurred or was
made necessary. Consequently, the subject may not be able to
recall his experience when asked. IntrosPective methods have also
shown that the subject is frequently unable to verbalize or to
describe his experience; hence, this method does not offer any
objective evidence that reformulation really did occur or, if so,
what the nature of the process was.

A more desirable technique of assessing reformulation is one
employing an objective index that can be obtained from direct
observation of the subjects' behavior at the point where reformu-
lation is necessitated. Such a method involves the establishment of
a task-induced set and the subsequent presenting of problems which
cannot be solved by that set. If it is assumed that the subject's
formulation of the problem determines what kind of solutions he
will get set to look for, failure will lead to reformulation and a
different set. Since formulation occurs during preparation, return-
ing to preparation can be taken as a Sign of reformulation.

Johnson has devised an apparatus which permits the separate
exposure and timing of two parts of a problem, "preparation" and
"solution. " During the preparation period, the subject is given
enough information to "collect his thoughts, " that is, to form a set
or a direction, towards the problem's solution. During the solution
period, the subject is required to make a reSponse which Solves

the problem. A toggle switch on the apparatus permits the subject

to see either side of the problem at will but not both. If the subject
finds on switching to the solution side of a problem that he is not
prepared to solve that problem (i. e. is set wrong) he can "switch
back" to the preparation side and re-examine the preparatory
parts of the problem. A "switchback" is taken as an index of
reformulation.

Johnson (14) obtained an objective index of reformulation
using this procedure and apparatus. As the present study is modeled
largely after Johnson's unpublished one, his experiment must be
brieﬂy summarized in order to adequately interpret the results of
this study.

Johnson used a series of eight training cards to build up task-
induced sets in his subjects. On each half of a card ten geometric
figures were diSplayed in various Spatial orientations. The subject' s
task was to decide what the ten figures on the preparation side had
in common. He then turned the toggle switch to light up the solution
side of the card on which ten more figures were diSplayed. Subjects
solved the problems by choosing the one figure on the solution side
that had the same characteristic as the figures on the preparation
side.

For one series of cards, the common characteristic was the

shape of the figures. For example, all of the figures on the prepara-

 

tion side were rectangles, or circles or triangles etc. For the other
series, the common characteristic was the interior shading or
texture of the figures. Each of the figures had, for example, vertical
lines, horizontal lines, cross-hatched lines or the like.

The effectiveness of the training series in establishing sets was
tested by presenting an "ambiguous" (cf. 23) situation to the subjects.

The display of figures on the preparation side had both shape and

ljullil '.

 

 

10

texture in common. The subject could solve these ambiguous prob-
lems by choosing either a shape or a texture reSponse. The highly
significant number of responses consistent with the training series
showed that the materials had established highly effective sets.
Hence the ambiguity of the situation was resolved by the set (cf. 31,
p. 830).

The objective index of reformulation was obtained from a
second group of subjects trained in the same manner. These sub-
jects were given test problems for which the established set was
inapprOpriate. The correct solution of their problems necessitated
a "breaking" of the established set and the formation Of the set
different from the one on which they had been trained. The index
of reformulation was obtained by counting the number of times it
was necessary for the subject to return (i. e. turn the toggle switch
back) to the preparation material before he could solve the problem.

The number of switchbacks made by this group was significantly
higher than the number made by both a control group who solved
problems consistent with their training sets and the ambiguous group.
Significantly more individuals made switchbacks than in the other
groups.

That the switchbacks did, in fact, constitute an index of
reformulation was shown by the re sults Obtained with a group trained
and tested in the same manner but not allowed to switch back either
during training or test trials. Significantly more errors were made

by more individuals in this group than in any of the other groups.

EXPERIMENT

Hypothe sis

 

Johnson's results showed that this method with its switch-
back procedure could be used to separate formulation from solution
and to obtain an objective index of reformulation without the un-
certainties of subjective observation. An implied conclusion from
his results is that an inapprOpriate set is one of the conditions
leading to reformulation. A question arose as to the generality of
the manual switchback as an indicator of reformulation in Similar
problems involving different kinds of materials. Specifically, could
the switchback index of reformulation be obtained with problems
employing verbal materials to induce sets rather than figures?

The hypothesis tested in the present experiment was that this
behavioral index of reformulation would be unaffected by the nature
of the materials of the problem. i

The present eXperiment, devised to test this hypothesis, made
use of words grOuped in such a way as to enable the subject to form
concepts on the basis of the groupings. Two types of concepts were
employed in a manner analogous to Johnson's figures. (A more
complete discussion of the concepts is presented at a later point

in the paper.)

Design

The subjects were ninety-7 six under graduate psychology students
drawn from classes during the Spring and summer terms of 1959.
They were divided into four groups of 24 each. Group I, the ambiguous

group, provided a test of the effectiveness of the materials in inducing

ll

12

sets. Group II, the reformulation group, was to provide the
switchback index. Groups III and IV were control groups. Group III
controlled for the difficulty of the materials and was to provide
evidence that factors other than the materials themselves were
reSponsible for switchbacks. Group IV was also a reformulation
group but they were not allowed to switch back. Greater errors in
this group would show that the switchback did in fact lead to
reformulation and subsequently to correct solutions. Each of the
four groups was further divided into two subgroups according to

the concept to which they were exposed during the training series.

Apparatus

 

The apparatus used in the experiment consists of a box divided
into approximately equal halves and equipped with a holder at the
back into which 5" by 8" index cards can be fitted manually by the
experimenter. Each half can be lighted separately and timed
separately by means of standard electric timers which are operated
by each circuit. For this experiment the timers were equipped with
Specially constructed logarithmic dials so that time scores could
be read directly in log units, thus precluding the necessity for later
conversion.

The left side of the box presents that half of a card on which
the parts of the problem essential to "preparation" are diSplayed.
The right side presents that half containing the possible "solutions"
to the problem. A one-way glass prevents the subject from seeing
the materials before the light is turned on in that compartment.
Stray light which would enable the subject to see through the glass is

controlled by conducting the experiments in semi-darkness. A small

 

l3

desk lamp sits under the apparatus and provides enough light to
read the timers.

E manipulates a master switch which simultaneously turns on
the light on the preparation side and activates an electric timer.
The subject studies the left side of the card as long as he wishes
and turns the toggle switch. This action darkens the preparation
side and stOps that timer while simultaneously illuminating the
solution side and activating this timer. The subject indicates his
solutions by pressing a numbered button on a bank in front of him
thereby darkening the solution light and stopping the timer that
measures solution time. The ﬂoor plan of the apparatus Showing

the lighting and timers is shown in Figure 1.

Materials

 

All instructions and words used in the experiment were type—
written in lower case pica type on plain white index cards. Five
words, approximately centered and double-spaced, were typed on

each Side of an index card in the following manner:

dandelion 1. sugar
butte r 2 . yell ow
gold 3 . throb
mustard 4 . brown
canary 5. scooter

The words on the preparation side were such as to enable the
subject to form a concept which solved the problem (in this example,
yellow). The remaining words were made to be as neutral as
possible.

Two types of concepts were used in the experiment. One type

was based on the sensory impressions (cf 24, 25) made by objects

l4

 

v w._ .,_ ._.._.,_ _. .., _ . Y .0- -. .
- ’v _ _ "‘1“.- "‘]h.'4.-a.p- ‘.- ’-L'*‘”. " '-‘ . . 3-.¢A—. .7
_ '- — ..- —‘ -' —~- "~)~... -—
v- ,V 5"" int-.- -: , -ﬂ

1; WM: ' .

 

 

 

 

 

~ .r‘i" :vv-t.

 

 

 

 

 

 

 

 

O
NO
W)

O

 

 

Figure 1. Floor plan of apparatus.

15

represented by words. An example of this concept appears above.
A preliminary study indicated that the practice of naming the con-
cept (yellow) led to the establishment of more effective sets than
did choosing another example of the concept (for example, corn).
Most of the words used for the sensory concept were taken from
Underwood and Richardson's list of high reSponse categories of
sensory impression (24).

The other type of concept was based on more abstract character-
istics. The words used for this concept could be subsumed under

such concepts as bodies of water, articles of furniture, fish, etc.

Procedure

 

All subjects read a short paragraph identifying the experiment
as one in "forming concepts on the basis of common characteristics
shared by the objects represented by words. " Subjects read that it
would help them if they tried to "form a global impression of how
the objects taken as a group would appear together. " This statement
was intended to preclude attempts by the subjects to analyze all of
the properties of any one object thereby ignoring communality with
the other objects.

Further instructions regarding the task and the apparatus were
presented in the box. On the left (preparation) side, subjects read
that they would next see a group of words which were all similar in
some way. They were to decide what these words had in common
and to remember the common element as they would have to find
another example on the other (i. 6. solution) side of the card.
Subjects read that they could turn the switch whenever they. felt that
they knew what the common element was. They were warned that the

left side would become dark and they would see only the right side

16

of the card. The last line of instructions, "You may switch back
if necessary, " was deleted from the instructions for Group IV
(See Appendix).

Subjects then turned the toggle switch lighting up the solu-
tion Side of the box where they read that they would next see five
numbered alternatives. They were to pick the one that was
another example of the other side and indicate their choice by
pressing the correSponding button on the board in frontaof them.

All subjects were then shown three practice cards to acquaint
them with their task. One of the sample cards, for example, had
the words, republic, representatives, congress, legislate and
elect typed on the preparation side. Subjects should have prepared
to solve this problem by determining that all of these words were
"political terms" (21). Consequently, on turning to the solution
side and seeing the words: .1. imagination, 2. amendments,

3. portrait, 4. conclusion, and 5. helicopter, subjects should have
been "set" to pick the word "amendments, " another political term,
as the correct solution.

After the three practice cards, subjects went directly into
the body of the experiment. Subjects were run individually in a
fixed order so that any population changes would be reﬂected in all
groups. No two consecutive subjects were given the same concept. ’
Ten training cards and five test cards were shown to each subject.
Partial randomization of the order of presentation was accomplished
by a shuffling procedure.

The training of Groups I, II and III was identical. Verbal
reinforcement was used to enhance the establishment of set and

switchbacks were permitted at any time. Group IV was trained in

l7

essentially the same manner except that this group was never
allowed to switch back. It was eXpected that the standardized
pretraining would provide uniform sets. Consequently, different
”test conditions” afforded the test of the hypothesis.

Group I was tested on five "ambiguous" problems which could
have been solved by either concept. For example, one of the
ambiguous cards contained the names of five "green vegetables"
on the preparation side. On the solution side a subject set for
sensory dimensions would find the word "green, " while a subject
set for abstract concepts would find the word "tomato. ” It was
expected that solutions would be made on the basis of the re Spective
training concepts if the words in the training series had been
effective in establishing sets.

Group II, the reformulation group, was tested on problems
for which the training set was inappropriate. Subjects who were
trained to use abstract concepts now had to reformulate the problem
in terms of sensory concepts in order to solve the problem. For
the green vegetable card, this group could now solve the problem
only on the basis of the concept "green. " The abstract-concept
group were similarly required to formulate the other concept to
solve their problems. The number of times subjects in Group II
switched back to the preparation side before they solved the test
problems is the objective index of reformulation.

The test cards for .Group (III were the same cards that were
used to assess reformulation in Group II. The cards were used
differently however in that the problems were consistent with the
subjects' pretraining. Since the established set would be apprOpriate,
it was expected that both errors and switchbacks would be minimal

for this group. If the number of switchbacks were significantly

18

smaller than for the reformulation group, evidence would be obtained
to support a conclusion that an inappropriate set is one of the con-
ditions necessitating reformulation and that the switchback is an
index of this reformulation.

Group IV subjects were treated in the same manner as Group
11 subjects in that the established set was inappropriate for solving

the test problems. It was assumed that if they could not reformulate

they would make errors.

RESULTS

The effectiveness of the training series in establishing sets
was determined by comparing the number of sensory choices made
in the ambiguous situation by each of the subgroups of Group I. ‘
As the two subgroups had had different pretraining, a significantly
different number‘of sensory choices by the subgroups could be
attributed to the type of pretraining that they had; in other words,
it would afford evidence that the pretraining had resulted in set
establishment. The frequency distribution of sensory choices

according to type of pretraining is given in Table I.

TABLE I -

FREQUENCY DISTRIBUTION OF SENSORY CHOICES
BY TYPE OF PRETRAINING

 

 

 

 

 

Pretraining Frequency of Sensory Choices __
0 l 2 3 4 5 3?

Sensory (n = 12) l 4 3 4 3. 83

Abstract (11 = 12) 4 3 2 2 l 1.42

 

3: 4.82; P< .001.

Inspection shows that the subgroup trained on the sensory
concept made a far greater number of sensory choices than did
the one trained on the abstract concept. The probability of such an
occurrence is less than . 001. Scoring responses on the basis of
the number of abstract choices yields complementary results.

Hence, it can be concluded that sets, or prediSpositions to make

19

U

( \.

 

 

 

20

reSponses consistent with those required in the training series,
were established by the training materials.

Since the pretraining did result in sets, it was expected that
giving a problem for which the established set was inappropriate would
lead to more returns (i. e. switchbacks) to the preparation material.
Consequently, Group II was expected to make significantly more
switchbacks than any of the other groups. This hypothesis was not
confirmed. The total nutnber of switchbacks over the five test trials
were 17, 19 and 15 for Groups I, 11 and 111 reSpectively (Table 11).
These values did not differ significantly.

TABLE II

ERRORS, SWITCHBACKS AND MEAN TIMES IN SECONDS
OVER FIVE TEST TRIALS

 

~ Preparation Solution
Group Switchbacks Errors Time Time
I 1 7 3 4. 3 1 4. 44
II 1 9 9 4. 12 4. 89
IH 15 3 3. 55 3. 64
IV - 10 4. 89 5. 75

 

Group IV, which brought an inapprOpriate set into a problem
situation which did not allow for any (apparent) reformulation or
restructuring of the situation, was expected to make more errors
than the other groups. However, this group made errors equivalent
to the number made by Group II who were allowed to reformulate

by switching back. The total number of errors made by each of the

21

groups is presented in Table II. It should be added that choices

by Group I were not considered errors if they were consistent with
either dimension of classification, abstract or sensory. No corre-
lation between errors and time scores were found. Average
preparation and solution times over the five test trials are also
given in Table II. Q

Number of appropriate solutions is but one way of assessing
the effectiveness of set (23). A second way is in terms of the
amount of time spent on a problem as an index of the efficiency of
the set. It follows that if a set is highly effective, the choice of a
reSponse in a problem situation for which the set is appropriate
will be in accordance with the set and will be made quickly.
However, if conﬂicting sets, or determining tendencies of some
kind, are operating, changes in time scores reﬂect this conﬂict
and consequently bear on set efficiency.

As switchbacks did not provide an index of reformulation,
differences in time scores between the training and test series were
analyzed not only to provide an alternate means of assessing the
effectiveness of the training sets but also in an attempt to determine
the implications of these time score changes for the reformulation
process. The method used involved calculation of the differences in
time scores, expressed in logarithmic units, between the last
training trial and the first test trial for each group. For example,
Group I subjects Spent 630 log units in preparation on the last
training trial and 639 log units in preparation on the first test trial.
These figures represent a net gain of plus 9 units for the group,
an increase which is not significant. These time score changes are
presented numerically in TableIII and graphically in Figures 2, 3,

4, and 5. Only the scores on the first test card were used as it

22

was assumed that the greatest impact of a set change would be
reﬂected here. This assumption was borne out by the performance
curves. Only the last training trial scores were used because of
the belief that the Single last trial scores fairly well represented

asymptotic values (Figure 6).

TABLE III

TIME INCREASE (LOG UNITS) BETWEEN LAST TRAINING
TRIAL AND FIRST TEST TRIAL

f
M

 

 

 

 

 

Increase t
Group Preparation Solution Preparation Solution
I 9 24 1. 37 2. 42*
II 32 67 2.37%< 5. 11*
III - l l 2 -'~ 1 . 02 . 13
IV 12 59 l. 63 4. 13*

 

*
Significant at or beyond . 05 level.

10 log (100 x time in seconds).

23

spuooas u; sun;

2.3mm. “OOH.

wmm

‘P'r—i

O HO: .H mddﬁmn H.

mew
_e__.

mm“

 

OH

 

4 |< <

 

‘ S

I,

so 3.30m

I ‘0
I‘ll“ ” k
I

 

Goﬂumummonnm 1 w .

 

H QSOHOnumHOEH “OOH use mswsMMHH Home. 053. soﬂdaom pad soﬂdudmmnnm .N oudmﬁm

(spuooss 801) snutl

_24

spuooss 11; earn,

OH

ﬂute e39 33.5. message

..almysa m m a a: 4 a m w w H .mw

.8338

GOEMu amounm

 

HH adoumuuuwﬁmwnh. “OOH. use mswnmmnﬁ no“ mega. coﬂdﬁom was soﬁmhdmonnm .m onﬁmﬁm

 

 

(spuooas 301) emu,

25

spuooss a; suit;

3398 ”EOE 2.9.88 msﬂﬁduﬁ

Ewan-E... m. wee a. e... ._._.

 

m;

0H.

 

 

Goﬂdnmmounm

HHH msonOuumHmEH “OOH pad wcwsﬁmuh. .HOH mega“. GOESHom was coﬂmnmmmunm

4‘ 0.13th

J

 

L.
0‘

(spuooes BO'I) QUIIJ,

spuooas u; emu,

26

32MB pm on. m Haw.” H. magnum." .H.

wmmaoawwhomwm
P_r_________

 

OH.

 

4

>H QdOHOuudewHH. HOOP paw muwCMMHH. HOW maﬁa. Goﬂdﬁom was sowpmnmmounm

eossaom

 

s

coﬂmummonm

.m 0.34th

 

(spuooss 801) emu,

Time (Log Seconds)

 

Figure 6. Training Trials Performance for 48 Subjects

Having Each Concept

    
 
 

[Prepa ration

Sensory

\\ folution / \
\ / / \
\ ‘ ~ I \

  
    

\ I I
Soluti‘dn

-— — _-\

Abstract

Preparation

T rials

27

10

4:.
Time in Seconds

 

DISCU SSION

The data of Table I support the conclusion that the training
materials were highly effective in establishing sets for a particular
kind of resolution of the ambiguous situation. Both concepts were
effective in determining choices. Hence, with effective sets
established, giving problems not consistent with pretraining should
have provided information relevant to the reformulation process.

It was expected that a numerical index of reformulation based
on an objective, quantifiable measure of subject behavior could be
obtained by the methodology of the experiment. Set formation is
assumed to be an important factor in successful problem solving.
Set largely determines the aSpects of the stimulus situation that the
subject will attend to and directs the solution of the problem. Faced
with a problem for which an existing set is inappropriate, the subject
must reformulate the problem in such a way that a new, more
appropriate set leading to a solution is formed. The effectiveness
of a set may be measured in terms of the frequency and quickness
of appropriate solutions. The method employed in this experiment
allows for both kinds of measurement and also permits for a study
of the reformulation process.

Johnson (14) in his exPeriment with geometric figures obtained
an objective index of reformulation from the number of times sub-
jects turned a toggle switch back to the preparation side before they
could solve a problem for which a set established by training was
inappropriate. The present study, using highly familiar verbal

materials so arranged as to establish concept-sets, did not obtain

28

 

29

predicted differences in the switchback index of reformulation.
Several reasons, primarily in terms of the amount and kinds of
materials used in each experiment, can be given to account for
this failure. Among them are the Span of immediate memory, past
experience as it affects the nature of the task and other factors
relating to the effectiveness of the training sets including the
"intrinsic" nature of the task.

Differences in the amount of materials used to form the
re Spective concepts in each experiment may partly account for the
differences in results, namely, the failure in the present experiment
to obtain the objective index of reformulation. Probably one of the
most important reasons for failure to obtain differences in switch-
backs in this experiment is the Span of immediate memory (22).

This experiment employed only five relatively simple words during
the preparation period. Subjects could not only form concepts, or
perhaps rather concept sets, based on these five words but many
could also apparently remember all or enough of these familiar
words to preclude the necessity of "taking another look" at the
problem (i. e. by switching back manually). In J ohnson's experiment
ten figures were diSplayed on the preparation side of the card. To
retain ten detailed and unfamiliar figures in immediate memory
while searching for a solution may be an impossible task for most
subjects.

Essentially subjects in Johnson's experiment solved their
problems in the same manner as subjects in this experiment; namely,
by forming concepts (shape or texture). Johnson's subjects appar-
ently abstracted out the one aSpect, either Shape or texture common

to all of the figures. Retaining only this one aSpect in memory,

30

subjects could then switch to the solution side and solve the problem.
Having learned and retained only one relevant aspect of the problem,
Johnson's subjects were totally unprepared for a "set change" and
it was necessary for them to switch back and re-examine the parts
of the problem before they could solve it.

To solve problems based on the unfamiliar figures required a
greater amount of learning than the verbal materials. The kinds
of materials used in each experiment apparently interacted with the
amount variable and also possibly established concept sets of differ-
ing strengths and perhaps at different levels (9) in each experiment.
The past eXperience of the subjects as it affected familiarity with
the mate rials, and, in turn, the amount of learning during the train-
ing series interacted with the span of immediate memory to contribute
to the failure to obtain an index of reformulation.

In Johnson's experiment, the ten unfamiliar figures required
a high degree of attention and learning by the subject in order to form
an effective concept or set. The five relatively simple words used
in this experiment had undoubtedly been seen by the subjects hundreds
of times during their lives; consequently, the task of the present
experiment involved manipulative skills as much as or more so than
it did learning. That little actual learning occurred in the experiment
is shown by the learning curves in Figures 2, 3, 4, 5 and 6. Over
the ten training trials, subjects improved their performance by only
about two seconds. This would indicate the materials were already
highly learned and should be relatively easy to recall if the situation
demanded it (cf 25). Except for Group III there is little indication
that the subjects learned a set that facilitated response.

When the subject turned to the solution side and found that his

set was not apprOpriate he could reformulate the problem (i. e. assume

31

a new set or direction towards the problem) in one of two ways.

He could ”switch back" to the preparation period as some did or

he could utilize his memory to recall the preparation material as
others apparently did. Some of the subjects apparently interpreted
the set change as a challenge to their abilities as several verbalized
that they "almost had" to switch back. All of these factors would
serve to depress the total number of switchbacks.

As both experiments involved the attainment of concepts in
the establishment of sets, an attempt at an interpretation of the
results in terms of some of the demonstrated or suggested features
of concept formation would seem to be in order. Wohlwill (27) has
suggested that two different cognitive processes may be involved in
concept formation. One process is a perceptually determined

process of abstraction which consists of a selective response to a

 

given aSpect of a stimulus. The other process he refers to as

conceptualization which he considers as a process of "mediated

 

generalization and, as such, being primarily a function of the prior
experience and stage of mental development of the individual. "

The formation of concepts by this latter process is seen as being
accomplished by the activation of "reSponse sets" in the eXperimental
situation which were formed during the course of mental development
and experience.

Thus the formation of concepts in Johnson's experiment could
be interpreted in terms of a perceptually determined abstraction
process. In this experiment, concept formation could be interpreted
in terms of a "supra-perceptual" process of conceptualization.

Heidbreder suggests that concepts are more readily'attained
at a perceptual level than they are at an intellectual level (9, 10).

"Concepts are more readily attained as the contexts in which their

a»\

32

crucial features are presented permit--conceptualization at a more
perceptual rather than a more intellectual level" (9). The greater
the "thing quality" of an object (10), that is, the more concrete its
representation, then the more readily attainable is the concept.

Reinterpreting Heidbreder in terms of the two experiments
under discussion here, the suggested explanation is that the con-
ditions of J ohnson' s exPeriment kept the whole course of conceptual-
ization close to the perceptual level, whereas the verbal materials
took the course of conceptualization to a more intellectual level.

In other words, Johnson's subjects perhaps came to rely only on
their perceptions of the figures to guide them in solving problems
and made little or no attempt to integrate their perceptual reSponses
with other information. Consequently, when the condition arose
where a perceptual reSponse of the kind they had been making was
not apprOpriate and having no other information to guide them, it
was inevitable that they should switch back to the preparation side
where they could reformulate or restructure the problem.

The familiar verbal materials already well learned and inte-
grated allowing for conceptualization at a more intellectual level
also allowed for the accessibility of additional information when the
established set failed. Consequently, switchbacks were not always
needed in order to make correct solutions.

Rees and Israel (23) demonstrated that sets can function at
different levels of awareness. They found that their letter order
set Operated at an almost automatic, unconscious, or at least un-
verbalizable, level to direct the solutions of the ambiguous anagrams.
Their category set however was almost always accompanied by

conscious awareness. That the automatic set was more effective

33

than the conscious set was shown by the arousal of both kinds of
set in the same experiment.

Rees and Israel interpreted the greater effectiveness of the
letter order set over the category set to be a function of the intrinsic
nature of the task. They reasoned that anagram solution was primarily
a process of rearranging letters and secondarily a process of search-
ing for solution words, hence, letter order set should be more
effective. While this paper would not argue with the general finding
that the effectiveness of set is dependent on the nature of the task,
the kind of analysis being used here would go beyond their explanation
and would consider the nature of the materials used as well as the
task itself.

Essentially, their letter order task involved making perceptual
responses to the concrete structure of anagrams. Their category
set task was a more intellectual task in the sense that it activated
old responses and Old associations in the search for category solu-
tion words. It may be that the "blinding" effects of set are due to
too complete reliance on the perceptual apparatus alone at the expense
of integration and assimilation of the materials learned or being
learned. Consequently, when there is heavy reliance on perception
of concrete instances Of concepts, it becomes an easy task to
"trick" the subjects by merely changing what they perceive. The
automatic functioning of certain kinds of sets may be due to per-
ceptual factors whereas conscious awareness of the set may be due
more to intellectual factors.

Johnson's figures may have deveIOped more automatic sets
than the verbal materials developed. Consequently, the automatic

set would serve to more effectively "blind" the subjects to alternate

34

methods of solution (cf Luchin's Einstellung l6), and make switch-

 

backs a necessary aid to correct problem solution. The sets
established by the familiar verbal materials could have remained

at a more conscious level with at least partial awareness of their
Operation by the subjects. Consequently, in the ambiguous situation
the effectiveness of the conscious set would be lessened either in
terms of apprOpriate solutions or in terms of lengthening reSponse
times. This latter condition appears to have occurred.

A significant increase in solution time scores was shown for
Group I. This increase may have been due to an element of "doubt"
in the subjects' judgments which was created by at least partial
awareness of an alternative solution. It may be that at least some
of the subjects realized that an alternate solution was possible but
that they still made choices consistent with their sets as this is how
they had been solving the problems; but, apparently the set was not
as firm as the one established by Johnson's figures. Rees and
Israel (23) demonstrated that the establishment of antagonistic sets
would serve to lengthen reSponse time s as the presence of one set
tended to hinder the operation of the other. In this case, the
re Sponse sets brought into the experimental situation were probably
activated by the familiar material on the ambiguous cards and
interfered with the task-induced set to the extent of increasing time
scores.

Significant increases occurred in both preparation and solution
time scores for Group II, the reformulation group, and also in
solution time scores for Group IV which was also a reformulation
group but which was not allowed to use the switch back as-an aid
to reformulation. As previously stated the kind and amount of

materials in this experiment allowed for reformulation, defined as

35

a change in the establiShed set such that the correct solution of the
problem is attained through the process, to come about in either
of two ways. Subjects in Group II could reformulate the problem
either by switching back or by utilizing their powers of recall.
Subjects in Group IV could only reformulate on the solution side.
Both groups had equivalent error scores. Consequently, reformu-
lation appears to have been approximately equal for both groups.

At an earlier point in the paper, it was stated that reformu-
lation was contingent upon a return to preparation to re-examine
the parts of the problem and to assess cues that had not previously
been re Sponded to. The manual return to preparation and the
"visual search" by Group 11 subjects who could not or preferred not
to utilize memory to reformulate is seen as being responsible for
the increase in preparation time scores.

Solution time scores for Group II increased partly as a result
of inﬂation of the scores by the return switchback. The larger part
of the increase, however, was probably due to the subjects' use of
memory for purposes of reformulation. Faced with a problem for
which their established sets were inappropriate, many subjects in
Group II and almost all in Group IV apparently returned to preparation
by a symbolic process of memory rather than a manual process of
switching back. Thus solution time scores increased for these two
groups because of the recall process. Reformulation is assume d to
have occurred as a result of this recall because of the small number
of errors.

The failure to obtain an Objective index of reformulation based
on switchbacks by no means invalidates the method. It suggests
rather that highly learned verbal material may not be amenable to

study by this method--that the only index of reformulation that can be

36

obtained with this kind of mate rial is a temporal one.) It may be
possible to devise an experiment similar to the present one which
would overcome the difficulties of the memory span but the
selection of apprOpriate materials is extremely limited.

The separate timing of the preparation and solution periods
make it possible to investigate each process separately. Through
this method it is possible to search for factors which might contribute
to one or both phases. One technique would be through the corre-
lation of sub-tests of general screening tests with; each phase of
problem solving. Some tentative overtures have already been made
in this direction.

Another approach to problem solving by this method would be
to present different kinds of problems to subjects to see whether
time spent on preparation and/ or solution remains constant over
problems of approximately equal difficulty. Individual differences
in problem solving can also be assessed by this method.

Set has been used previously in investigations of concept
formation (26). The methodology of the present experiment allowing
for a study of set establishment also allows for a study of concept
formation. Given a problem for which an established set is in-
apprOpriate, the subject must essentially invoke or form a new
concept in order to solve the new problem. Using a series of such
problems, it may be possible to obtain information relevant to the
nature of concept formation by studying the separate reformulations

and analyzing time records of the subjects.

SUMMARY AND CONCLUSIONS

A method of studying problem solving processes was utilized
in an experiment designed to afford an Objective index of reformu-
lation based on quantifiable observable subject behavior. Two
concepts were used in a training series to establish task-induced
sets in the subjects'. The subjects were then given problems for
which the established set was inapprOpriate. It was expected that
the number of times they returned to the preparation part of a
problem by a switchback procedure would yield an objective index
of reformulation.

It was found that set was established as expected, but that the
predicted differences in reformulation did not appear. It is sug-
gested that time score increases are a sign of reformulation during
the solution period of a problem and that this reformulation was made
possible by the subjects' ability to recall the familiar verbal

mate rial of the problem obviating the necessity for switching back.

37

BIBLIOGRAPHY

l. Adamson, R. E. Functional fixedness as related to problem

10.

11.

12.

solving: a repetition of three experiments. J. exp. Psychol.,
1952, 44, 288-291.

 

Birch, H. G. and Rabinowitz, H. S. The negative effect of
previous experience on productive thinking. J. exp. Psychol. ,
1951, 41, 121-125.

 

Broadbent, D. E. A mechanical model for human attention and
immediate memory. Psychol. Rev., 1957, 64, 205-215.

 

Chapman, D. W. Relative effects of determinate and indeterminate

Aufgaben. Amer. J. Psychol., 1932, _4_4_, 163-174.

 

Dolinko, P. Set and conceptual defense. Unpublished Ph. D.
Dissertation, Purdue Univ. 1957. In Dissertation Abstracts,
1957, 11, 1388.

 

Duncker, K. Zur Psychologie des produktiven Denkens.
Berlin, Springer. Trans. L. S. Lees. On problem-solving.
Psychol. Monogr., 1945. No. 270.

 

 

Gagne, R. M. Problem solving and thinking. Annual Rev. of

 

Psychol., 1959, 10, 147-172.

Gibson, J. J. A critical review of the concept of set in contemp-
orary experimental psychology. Psychol. Bull. , 1941, 38,
781-817.

Heidbreder, E., Bensley, M. L. and Ivy, M. The attainment
of concepts: IV. Regularities and levels. J. Psychol., 1948,
_2_5_, 299-329.

 

 

Heidbreder, E. and Overstreet, P. The attainment of concepts:
V. Critical features and contexts. J. Psychol., 1948, 26, 45-69.

 

Hunter, I. M. L. The inﬂuence of mental set on problem solving.
Brit. J. Psychol., 1956, 41, 63-64.

 

Jensen, P. K. Differential Effects of Distraction on Thinking.
Unpublished Master's Thesis, East Lansing, Mich. , 1954.

38

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

39

Johnson, D. M. The Psychology of Thought and Judgment.
New York, Harpers, 1955.

 

Johnson, D. M. Serial Analyses of Thinking. Unpublished
paper.
Kiilpe, O. Versuche iiber Abstraction. In Ogden, R. M.

Oswald Kiilpe and the Wiirzburg School, Amer. J. Psychol. ,
1951, 64, 4-19.

 

 

Luchins, A. S. Mechanization in problem solving: The effect
of Einstellung. Psychol. Monogr., 1942, it, No. 248.

 

Maier, N. R. F. Reasoning in Humans. I. On direction.
J. comp. Psychol., 1930, _1_0, 115-143.

 

Maltzman, I. and Morrisett, L. Jr. Different strengths of
set in the solution of anagrams. J. exp. Psychol., 1952, i4,
242-246.

Maltzman, 1., Eisman, E., Broodk, L. O. and Smith, W. M.
Task instructions for anagrams following different task
instructions and training. J-. exp. Psychol., 1956, §_1_, 418-420.

 

 

Marvel, J. A. Acquisition and retention of reading performance
on two reSponse dimensions as related to "set" and tachistosc0pic
training. Unpublished Ph. D. Thesis, Univ. of Okla., 1955.

In Dissertation Abstracts, 1955, 1_5_, 2132.

 

Mayzner, M. S. and Tresselt, M. E. Shifts in Connotative
Meaning of Words as a Function of Varying Amounts of

Previous Restrictive Experience. J. of Psych., 1959, 47,
107-116. _

 

Miller, G. A. The magical number seven, plus or minus two:
Some limits on our capacity for processing information.
Psychol. Rev. , 1956, _6_3_.

 

Rees, H. J. and Israel, H. E. An investigation of the estab-
lishment and operation of mental sets. Psychol. Mannogr. ,
1935, :l_6., NO. 210.

 

Underwood, B. J. and Richardson, J. Some verbal materials
for the study of concept formation. Psychol. Bull. , 1956, _5_3_,
84-950 '

 

Underwood, B. J. and Richardson, J. Verbal concept formation
as a function of instructions and dominance levels. J. exp.

Psychol., 1956, _5_1, 229-238.

26.

27.

28.

29.

30.

40

Vinacke, W. E. The investigation of concept formation.
Psychol. Bull., 1951, 18, 1-31.

 

Wohlwill, J. F. The abstraction and conceptualization of form,
color and number. J. exp. Psychol., 1957, _5_3_, 304-309.

 

Woodworth, R. S. Experimental Psychology. New York,
Holt, 1938.

 

Woodworth, R. S. Contemporary Schools of Psychology.
New York, Ronald Press, 1948.

 

Woodworth, R. S. and Schlosberg, H. Experimental Psychology,
New York, Holt, 1956.

 

APPENDIX

41

MATERIALS

The following instructions, printed on an index card, were
shown to all subjects after they were seated:

"This is an experiment in forming concepts on the basis
of common characteristics shared by the objects repre-
sented by words.

It will help you if, during the experiment, you try to form
a global impression of how the objects, taken as a group,
would appear together. "

The following instructions were placed in the box for all
subjects to read:

On this side of the box, you will On this side of the box, you will

next see a group of words which next see five numbered alterna-
are all similar in some way. tives. Pick the one that is

You are to decide what these words another example of the other side.
have in common. The common Indicate your choice by pressing
element will be easy to find in the correSponding button on the
others. After you have discovered board in front of you. HOLD the
what the common aspect is, button down until the exPerimenter
remember it as you will have to says 0. K. Then release the

find another example on the other button and turn the switch back to
side of the card. its original position.

When you feel that you know what Now press any number on the

is common, turn the switch on board to Show that you have
the box to light up the other side. finished reading the instructions
This side will become dark and and we will begin.

you will see only the right side.
(You may switch back if
necessary).

Now turn the switch.

42

SAMPLE CARDS

l. 2.
five 1. hate republic 1. imagination
eight 2. miss representatives _2_. amendments
seven 3. toy congress 3. portrait
four _4_. six legislate 4. conclusion
nine 5. junk elect 5. heliCOpter
3.

science 1. minister

laws 2. sonnet

proof 3. heaven

results 4. bullet

evidence 2. conclusions

- indicates a correct answer.

SENSORY SET

enaerald
lavni

ivy

dollar bill
cucuxnbe r

glacier
iceberg
snout
icecube
december

globe
wheel
barrel
donna
ball

atom
cap sule
ge rm
village
gnat

chalk
milk
sugar
te eth
ivory

rock
hail stone
helmet
stone
iron

ammonia
garlic
garbage
cigar
sewe r

dandelion
butte r
gold
mustard
canacy

H mph-combs merNH mlrs-wNu-alm'usoomw mpmNIr-lmpprp-a 01]th
' O

43

1 . e rg
tweezers
braid
Spinach
scamp

. brim stone
tongue
coral
tape

(c old

head
crash
roon1
crazy
dog

noon
trap
Inad
lﬁnge
snmdl

table
sma sh
curse
white
sing

vdndovr
house
hard
nest
hope

skunk

. able
scatte r
trail

. 1 brain

sugar
. yellow
3. throb
4. brown
5. scooter

IN

c ontinued

SENSORY SET - Continued

doughnut
saucer
lnﬂion
basebaH.
spool

apple
beet

blood
blush

cranberry

44

fast
circular
book
door
table

crab
soup
red
fancy
glade

ABSTRACT SET

hammer
saw

pliers
screwdriver
drill

zipper
button
cla Sp
hook
pin

perch
bass
pike
carp
t rout

river
ocean
streann
'lake
bay

umbrella
boots
hat
slicker
raincoat

chest
bed
sofa
table
lamp

c harleston
fox trot
waltz

cha cha
tango-

mph-cowl»- |U1u>w~H maswlwl-a mlmmww U'IVPUOINH mptszh-a U'Iv-FIUONH

45

budget
nod
wrench
cape
madam

snap
baker
udng
raid
cave

melon
tuna
caper
lamp
stove

birch
jewel
meat
pond
bowl

rose
rubbers
Whittle
paint
nail
course
ﬂour
cable

pepper
rug

rhumba

bottle
match
silly
casket

continued

ABSTRACT SET - Continued

potatoe s
Ineat
rice
bread
soup

airc raft carrier
de stroye r
gunboat

battle ship
submarine

fire

torch

lantern

lamp

ﬂare

O

lmrhmwr-a U1)rtht-I mn-Mri-I

46

porch
cup
eggs
board
ﬂoor

home
plant
cinde r
c rui ser
powe r

c ricket
bottle
silly
vauh:
match

aSparagus
lettuc e
cucumbe r
Spinach
peas

cherry
apple
tonnato
beet
cranberry

london
Chicago
be rlin
tokyo
mo scow

circumference
cylinder
Sphere

circle

arc

sulphur
formaldehyde
ether

chlo rofo rm
ammonia

UII-P-IUJNI-a
.

Ambi guou s

 

. biscuit
vehicle
green
truck
tomato

|U1 11-»le H

problem
blood
cushion
banana
key

U1Irl> PIN 1—4

diamond
holt

me rcy
ho spital
large

Iml-P-UOINI—a

round
baby
hexagon
plant
reap

carbon
night
clock
smelly
money

mlukowtvIn-i

"TEST CARDS"

Sensory
Refo rmulation

 

drawer
. vehicle
biscuit
. truck

. tomato

lmvthh-n
O

. problem
bush
cushion
banana
key

cum-cow:-
O O

p...»
o .

diamond
holt
mercy
ho Spital
bored

1.11»?th

plant
baby
hexagon
train
reap

UHF-Immu-
0

carbon
night
clock
ribbon
money

Ulv-F-DJNII—I
O

47

Abstract
Refo rmulation

 

biscuit
vehicle
green
truck
snore

mph-Icon)»-
0

problem
blood
cushion
bottle
key

01erle
0

diamond
break
me rcy
ho Spital
large

lU'll-PUJNI—l
O

. round
baby
plant
help
reap

UlrbUONIh-I
O

. clo set
night
clock
smelly
money

UlII-P-UONt-t
O

31013511 USE Gi‘iLY

 

MICHIGAN STATE UNIVERSITY LIB A I

RRES

3 1293 03062 1746