A SOCIO'LINGUISTIC STUDY OF
ELABORATED AND RESTRICTED
CODE QYSTEMS

Thesis for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
DUANE DENNIS PEITERSEN

1970

 

hm

 

\‘HESIS y 3‘ E: .. “i r»
' * L'Ak 11 R 1’ 2

OP

mic-i1: gran State E"

University

 
 

This is to certify that the

thesis entitled

A SOCIOLINGUISTIC STUDY OF
ELABORATED AND RESTRICTED
CODE SYSTEMS

presented by
Duane Dennis Pettersen

has been accepted towards fulfillment
of the requirements for

Doctor of Philosophy Jiegreein Communication

Major professé

November 18, }970
Date ‘

 

0-7639

 

 

 

.. ._.._.__._.___‘_...

A41 . __W

 

 

 

 

 

A SOCIOLINGUISTIC STUDY OF
ELABORATED AND RESTRICTED

CODE SYSTEMS

BY

Duane Dennis Pettersen

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Communication

1970

GTE/M

Accepted by the faculty of the Department
of Communication, College of Communication Arts,
Michigan State University, in partial fulfillment
of the requirements for the Doctor of PhilosoPhy

degree.

film; I? (firms/ZN—

Director of ThesisV

 

Céiafllézaaflfgaey Zé:7, Chairman
/'

ABSTRACT
A SOCIOLINGUISTIC STUDY OF

ELABORATED AND RESTRICTED
CODE SYSTEMS

by Duane Dennis Pettersen

The research reported in this paper concerns the
sociolinguistic theory of restricted and elaborated code
systems. The two linguistic codes are identified with
lower-working class families and middle class families and
specifically with family interaction patterns within the
social classes. The two codes are distinguished on a linguis-
tic, syntactic and lexical,level as well as a psychological
level. Linguistically, a restricted code selects a smaller
set of syntactic and lexical options than does an elaborated
code user. Psycholoqically an elaborated code-user expresses
a greater degree of "intent" or specificity than does a
restricted code-user. The two code systems result in differences
in information processing abilities, thus affecting users of
the codes in quite different ways in terms of expressing and
reacting to identical phenomena (social, intellectual and
emotional objects).

The sociolinguistic theory of Bernstein suggests that
working class teenagers will exhibit a restricted code system

and that a middle class teenager will employ an elaborated

_. a”

Duane Dennis Pettersen

code. Further it is predicted that teenagers will employ a
linguistic code more similar to their mothers than to the
code system of their peer group from a different social
class. William Labov predicts that teens will exhibit a code
system more similar to their peers than to their mothers.

Family discussions, eight families from each social
class, of relevant teenage issues were taped and submitted
for content analysis of ten variables. Working class members
were predicted to use a greater proportion of total pronouns/
total words, pronouns you and they/total pronouns, socio-
centric sequences/total words, and Taylor Cloze Procedure.
Middle class family members were predicted to use a greater
proportion of pronoun I/total pronouns, ego-centric sequences/
total words, subordinate clauses/total finite verbs, passive
verbs/total finite verbs, preposition of/of + in + into,
and the Gillie Abstraction Index.

Support was obtained for all hypotheses except the
proportion of pronouns/total words, and the Abstraction Index.
Thus significant differences were obtained between social
classes as regards a restricted and an elaborated linguistic
code system employed by working class and middle class
families, respectively. Further, the research supports
Bernstein's prediction that mothers and teenagers would employ

a more similar code system than teens and their peers.

Duane Dennis Pettersen

Implications of the research suggest further re-
finement of linguistic variables in terms of a generative
transformational grammar, and research of a longitudinal
nature focusing upon the effects of restricted and elaborated
codes on information processing abilities and academic

achievement.

ACKNOWLEDGMENTS

The writer wishes to express his appreciation to the
staff members of the Department of Communication, Michigan
State University. Special appreciation is extended to my
advisor Professor Erwin P. Bettinghaus who has been of in—
valuable service throughout my graduate program and completion
of this research paper. Thanks are also extended to Chairman
David K. Berlo and Dr. Randall P. Harrison of the Communication
Department, for their inputs in teaching and research. For
guidance during stages of the dissertation, I wish to thank
Professors Gerald R. Miller, Larry Sarbaugh and Hal Hepler.

To my colleague Dr. Daniel E. Costello, an extended appreciation
for his friendship and for his assistance during the data
collection stage of the dissertation.

Special consideration is given to my wife, Shirley, who
has assisted me in innumerable ways throughout graduate school
and finally as my faithful typist, proofreader and therapist,

while completing the dissertation.

".r._ A“

 

M1

TABLE OF CONTLITS

II‘TI-RODUCIF IO I o o o o o o o o o o o o o o

CPAPTEP

I SOCIOLINGUISTIC TFEOPY OW ELAPORATED

RESTRICTED CODES . . . . . . . .

General Sociological Theory
Information Frocessinr . . .
Labov's fiheory . . . . . . .
Summary . . . . . . . . . .

II RESEARCH LITERATURE AID PVFOTIFSFS

Fresummarv . . . . . .

O O O 0
fxnerimental Research . . . .
Bernstein's Pesearch . . . . .
Lawton's Pesearch . . . . . .
Robinson's Desearch . . . . .
Pesearch Areas and lynotheses
Summary . . . . . . . . . .

III RESEARCH DPSICD . . . . . . . .

General Study Design . . . . .
Sampling Procedure . . . . . .

Definitions of Dependent Var ables

Data Collection . . . . . . .
Data Analysis . . . . . . . .

N p PFULrI‘S O O O O O O O O C C O O
V COICLUSIONS . . . . . . . . . .
V I ‘C’UI’TVA R Y O O O O O O O O O O O O

PEPE? EP-CES . O O O I O O O O O C O O O 0

iii

AID

Page

 

 

 

 

A..
A;

 

 

LIST OF TABLES
TABLE:

l.--Descriptive Characteristics of Ibrkinq and
Middle C1888 Famnles o 0 Q o o o o o o o o o o

2.--Fendall Coefficient of Concordance: h
Analyses of Eight Linguistic Variables
Across Four Crouns, IC Mothers and Teenagers
and NC Mothers and Teenagers . . . . . . . . .

3.--Combined Scores of Pbrkina Class “others and
Teenagers vs. Combined Scores for Tfiddle
Class ”others and Veenacers as Tested bv
the Uann-Ihitney U Statistic . . . . . . . . .

U.—-Fruskal-Iallis one-wav Analysis of Variance
Among Four Groups (WC and WC Mothers and
Teenagers) on Each of Fight Variables . . . .

5.--*ann-thitney U r“est oP Significant Differences
between Pairs of Pbrkin: and Middle Class
I'Others and weer-19.0.8148 . g 0 0 g Q Q o o o o o .

6.--Differences between Ybrkinc Class and
Middle Class Family Discussions as Veasured
by Taylor's Cloze Procedure with Use of
the t-test . . . . . . . . . . . . . . . . . .

Page

 

56

60

7O

71

7d

77

.....

.......

............

oooooooooo

 

 

LIST OF CHARTS

 

Page
CHART l.--Summary of ranks assigned horizontallv
to each of four groups in terms of
variable usage . . . . . . . . . . . . . . 76

CFABT 2.--Croun ranking in terms of usage of eight
linguistic variables . . . . . . . . . . . 86

LI.“T 013-1 A PF} NI IC PS

Fa we

 

APPPIDIX A: LITTER FIIT TO RESPONDEITS , . . 109
APPENDIX E; LIST on DISCUSSION nUESTIors . . 1,;
APPENDIX c: FUISTIOIhATPPS . . . . . 0

APPENDIX D: IrSTPUCTIors AND SauPLE‘MPFSAnP‘ ' . 106

FOR cornucmrwn mnYLn.
.J J _ _._ _ 77 . . 7/ F, ' S I "I ‘
PDOCIDURP_ . . . CLOLI
o o o o o o o o o . . . l l 5

vi

3':
393“ F.‘
Qh‘b!
a...‘ :15

. ‘L1

as:
6
S

\t“

 

 

sad

A.

1!!
I'll.

INTRODUCTION

Culture is transmitted
largely through the medium
of lanouace, and behavior
is in large measure both
learned and expressed
through language.

Dell Hymes (1962)

The present research is based upon a concern for
sociolinguistics, a recent term created by social scientists
who have recognized the need for a better understanding and
use of linguistics in behavioral science research. This
study was undertaken as a result of the experimenter's
interest in Basil Bernstein's sociolinguistic research
conducted at the University of London. Sociolinguist
William Labov of Columbia University was also of interest as
Labov's work suggested certain theoretical differences which
the present writer wished to examine more closely. flhese
differences center on whether the parents (Bernstein) or the
peer groups (Labov) have the greatest influence on a child's
language behavior.

The interests of both Bernstein and Labov center on
the correlations found between social class and more speci-
fically family interaction patterns within social classes,
and the kinds of linguistic codes used by children. Bernstein

has identified two linguistic codes, called elaborated and

restricted, which seem to be determined by, if not correlated

ID

with, a middle class environment and a working class environ-
ment, respectively. The two codes are generally distinguished
by the degree of explicitness or intent expressed in the
linguistic code. Flaborated codes express a greater degree
of intent than do restricted codes. The codes are further
distinguished linguistically by the probability of use of
structural and lexical options.

The interest of the present research is twofold:
(l) to attempt to replicate in the United States, some of
Bernstein's research with the addition of testing mother
and child in the same environment thus providing further
clarification of Bernstein's sociolinguistic theory; and
(2) to provide a critical test of the conceptual difference
between Bernstein and Labov. The general procedure con-
sisted of interviewing a sample of middle and working class
families whose teenage sons attended a common high school.
Discussions by a 15 year-old teenager and his two parents
were tape-recorded and subjected to linguistic analyses.

lhe present research is organized into the following
chapters: Chapter I will examine current sociolinguistic
theory; Chapter II reviews related experimental research
and develops specific hypotheses for the present study;
Chapter III provides an explication of the study's design,
sampling procedures, data collection and statistical analyses;
Chapter IV presents the results; Chapter V discusses general

and specific conclusions and implications for future theory

building and research endeavors; and finally, Chapter VI
provides a brief summary of the research reported in this

document.

 

J

< ’u. P. A
C ..:
. r” «T
.. 3. a.
s I. . fl
A. n.
6.. “I

«LC.
.3 l

CIADIPV I

HUCIOLJW’HJISTIC TIIIWjV OV‘IUQAFOBACITJ AID PIIflfliICTPIJ(XNIfS

General Sociolinguistic Theory

 

One of the most important movements in behavioral
science since Iorld Par TI is interest in the study of basic
processes of communication and their regulative functions.
Study of the educationally disadvantaged has also led to an
attention to the process of language acouisition, to the
irelationships between language and cognition and to the
social antecedents and regulative conseouences of forms of
language use.

The significance of the role of language in cultural
development, argue Fereiter and Englemann (1066), is that
'cultural deprivation' may be defined as language deprivation,
especially for educational purposes. Language deprivation
is mainly a failure to learn the uses of language. nhe
problem of the culturally deprived, or the educationally
disadvantaged, the authors suggest, concerns one's ability
to obtain and transmit information, i.e., using language as
the primary information processing system for data manipu-
lations. Fereiter (1065) suggests further that information
processing abilities are the essence of intellectual functioning.

The suggested notion of differences in language-use or
using language as an information processing medium, was

recently given empirical support in Britain by Fasil

Bernstein (1958) in a discussion of elaborated and restricted
linguistic codes as found in general use by middle and
working class groups respectively. Bernstein states that

a function of language is that it "exists in relation to a
desire to eXpress and communicate; consequently, the mode

of a language structure--the way in which words and sentences
are related--reflects a particular form of the structuring
of feeling and so the very means of interaction and response
to the environment." (1958, p. 161) As Sapir (1956, p. 70)
states: "Language is heuristic...in the much more far
:reaching sense that its forms predetermine for us certain
Inodes of observation and interpretation." It is through
Specific linguistic codes that relevance is created, ex-
‘perience aiven a particular form, and social identity con-
strained.

Bernstein (1958) araues that changes in the form of
social control patterns act selectively upon the principles
of selection of both syntactic and lexical options. A
similar argument was advanced by Chomsky (1957) concerning
an innate structure which is universal for all men and which
provides or initiates a propensity for language acquisition.
The cultural or social milieu in which an individual is
born, determines the subset of rules, from a larger theoretic
set of rules, (syntactic, lexical and phonological rules)
which will be acquired, thus determining the particular

linguistic code employed in communication. Until recently,

relatively few social scientists were concerned with the
different subsets of rules that might be selected by groups
within a major linguistic code culture. That is, an English
code is an English code, and aside from various dialects

where pronounication or phonological differences were analyzed
and certain lexical idiosyncracies were recognized, no

serious effort was undertaken to find systematic structural
or grammatical differences.

Roger Shuy (1968) has discussed some of the tools
lavailable for evaluating as well as teaching nonstandard
sspeakers a standard English code system. He states in part:

Most current materials deal with pronun-
ciations although it has long been accepted
that grammatical differences count more
heavily toward social judgments than phono—
logical or lexical differences.

It stands to reason that there is a
hierarchy of importance in matters of teaching
standard [English] to nonstandard speakers....
If grammatical matters count more heavily in
social judgments, it seems reasonable to
assume that grammatical matters should receive
high priority in materials developed. (p. 83)

The implications of the use of two different lin-
guistic codes are significant in terms of the communication
activity between the two code—users. Identical phenomena
(social, intellectual and emotional objects) will likely
be perceived differently and thus relevant characteristics
of the object will differ as will interpretations about the

object. A suggested reason is that the restricted code-user

pays more attention to the content of the object and the

elaborated code-user to the structural relationships of the
same object. One implication of this difference as

revealed in the linguistic code, may be seen in student-
teacher relationships. Because the restricted code-user

does not cenerally distinguish between the role and person
who fills the role (its content), there is a frequent 'communi-
cation breakdown'. Eor example, the student who uses a
restricted code may consider the teacher who uses an
elaborated code, verv impersonal and 'cool' because the
teacher employs a much less freouent use of nonverbal cues
xnhich are necessary for the restricted code-user to identify
in providing meanings. Similarly, the teacher may perceive
the student as aggressive, rude or hostile because of the
absence of identification of linguistic intent and Specifi-
cation in the restricted code of the student. The teacher
Inakes the assumption that status differences (a structured
object) between the student and himself should be identifiable
in the languace of the student. However, the restricted
code-user does not make such a distinction lincuistically.
Freouent situations of inappropriate use of lanmuace by the
working class child and misinterpretations of the child's
intent by teachers, or vice versa, result in inhibitinn
social interaction. This may lead to a misevaluation of each
other's world and communication may become a means of
asserting differences.

The situation described in the previous paragraph

 

«Q 4*

 

r. J . , A 5;
n. O. n“ w 1.. Sc 5 v 0.. a.» «C.
3 ... v. 2. +t a. 3 w. 7. 2..
r. . . v o. .. ... “a E T .J r.
2. s . .2 ma : ». a: we
3 I 5.. a. n n. p“ .C :i Q. ~..
n. a» C... Q; C 2. A. a: n. D»

 

creates a resistance to learn new vocabulary or to constructinc
ordered sentences. From the child's point of View there

is no need to specifically verbalize intent. Personal
qualification through expressive symbolism, e.q., the use

of 'I', is adequately communicated by tone-volume-physical

set, not in the language he uses. His total perceptual

system results in a sensitivity to the content of an object

(DP its parts rather than the structure of the object, which
iJiduces a sensitivity to relationships amonq parts of the
cflaject or between one object and another.

The focus upon content with concrete, tangible and
suimple descriptive statements, with a 'now-orientation',
rwesults in further problems. The school, as an institution,
ennphasizes the structural features of events or objects,
mthich might be used in the future for different and more
(nomplex means of handling data and thus perceiving one's
Guivironment. This means-end orientation is not a focus of
tune child who looks for immediate satisfactions. Piaget's
C1958) differences in concrete and structural operations in
CCHrnitive development appears to lend support to the hy-
EMDthesized Operations which restricted and elaborated
liJimuistic codes allow.

The research to be reported in this paper, discusses
tvno linguistic codes which are used by people representing
individuals from a heterogeneous social structure and/or
family environment. The differinq environments appear

basic to the emergence of two codes, designated 'restricted'

and 'elaborated'. According to the socio-linguistic theory
first postulated at the University of London by Basil
Bernstein (1958), the two codes give rise to somewhat
different ways of perceiving and manipulating phenomena,

and specifically give rise to differing levels of performance
and academic achievement in a formal school environment. To
complete the cycle, Bernstein succests that the codes developed
reinforce the existing social structure which gave rise to
tune codes. Bernstein does state that restricted codes are
rust necessarily linked to any given social class. Restricted
(nodes are used at times by all members of a society. The
Inajor function of this code is to define and reinforce the
form of the social relationship by restricting the verbal
signalling of individual experience. A restricted code

<does not necessarily affect the amount of Speech, only its
form. anmples of various uses of a restricted code at times
are groups such as: husband-wife; a small prison proup;

or any closely knit group which has shared many common ex-
Deriences together.

Before continuing a discussion of the nature of the
tvn: codes and the social environments which purport to in-
f‘luence the develOpment and use of the codes, it would seem
Llseful and perhaps more meaningful to first look at language
arui language codes in terms of an 'information processing
franmwork' and particularly in terms of an analogy to the

function of a machine language in computer processing.

10

Information Processing

 

The framework of information processing for a computer
model consists of three basic components: an input-output
system, a processing or programming component, and a storage
or memory unit. The major concern here is the processing or
programming component. In order to utilize the data input,
the computer or the individual 'information processor' must
have the basic 'program'. The program of the computer is
transmitted through the use of machine language or a symbol
system for 'telling' the machine what to do. The language
of the program 'tells‘ the computer what information to
accept, i.e., determines the input, where it is to be
stored, what operations are to be performed, what order the
Operations are to follow, and what output is emitted.

The analogy of the computer to human problem-solving,
decision making and thinking, in other words human information
processing, was suggested more than a decade ago by
H. A. Simon of Carnegie Institute of Technology and A. Newell
and J. C. Shaw of the RAND Corporation (1958). More recently
James Miller (1965) has viewed language as a special sort
of human artifact, an information-processing artifact.
Language, as with music, and other information trans-
Inission mediums, is designed to carry out some critical
pmocesses essential to a living system. Whether it is a
Iiatural language of man or a machine language of some computer
system, Miller states that language is essential to infor—

Ination processing.

fir vv"
hue c ‘

L

‘l-‘np

L...»

A...

h.
9.
he

hic’

'7
I"

:;

S
Q»

a

IF H o I:
‘ c s. S ‘\'
.. ,

at

Up

11

As the machine language determines the processing
functions of the computer, so human language programs what
may be accepted, the interpretation of what is accepted,
the integration and memorization of data, and the output of
the system. Although language as a linguistic code is not
the only medium for program direction in the human information
processing unit, it seems to be the most important, especially
in terms of academic studies. How effectively and efficiently
the person functions is dependent on the 'program' that is
used for processing the data. According to Labov (1970)
one of the serious drawbacks in present linguistic knowledge,
is that we don't know what programs or aspects of language
actually help us to exnress our ideas clearly and facilitate
cognitive deve10pment. Thayer (1967) contends that the
notion of 'rules', 'policies' or 'programs' for information
preferencing and processing lies at the heart of any study of
human behavior. The problem is one of determining which rules,
which policies, and which programs of linguistic behavior

are significant.

Bernstein's Theory

 

Restricted and elaborated codes, or information-
processing programs, as described by Bernstein, are tenta-
tively interrelated with social class, but more Specifically
With the social structure of subgroups within a society, and

in particular the structural patterns within a family unit.

12

Bernstein suggests that two general types of codes may be
defined on a linguistic level in terms of the probability
of the use of linguistic structural and lexical Options.
That is they can be defined in terms of the probability of
predicting for any one speaker which structural elements will
be used to organize meaning. A restricted code is a less
informative code in that the kinds of structures employed
by a speaker are fewer or more severely constrained, and the
probability Of predicting the pattern Of organizing elements
is greatly increased. An elaborated linguistic code-user
will select from a relatively extensive range of alternatives
from the syntactic and lexical Options available in the
language system. In other words, the two codes might be
viewed as follows: given a theoretic set Of all possible
syntactic and lexicon Options in the English language system,
certain social-cultural environments seem to give rise to a
language system which systematically selects a smaller subset
Of the universal set. Specific features of the linguistic
system of the restricted code-user characterizes particular
styles and kinds Of role behaviors as well as perceptual and
meaning functions that are markedly divergent from those
behaviors emitted by the social-cultural environment Of the
elaborated code-user.

On a psychological level, Bernstein has distinguished
the codes in terms of the extent to which each facilitates
(elaborated code) or inhibits (restricted code) the orien-

tation to symbolize intent in a verbally eXplicit form.

13

When the intent of the other person cannot be taken for
granted, with the conseouence that meanings must be ex-
plicitly verbalized, verbal planning promotes a hioher level
of syntactic orcanization and lexical selection. The prepara—
titulzand delivery of relatively explicit meaning is the
major' function of the elaborated code. It is this construct,
intern: or greater or lesser explicitness, which is the over-
all cfliaracteristic of focus in Bernstein's theory. Exactly
what zaSpects of the elaborated code are critical for
sypflxolizing intent is not necessarily acreed upon by lin-
guistxic researchers. Bernstein (1966) suggests that whatever
are tflne critical features of an elaborated code, they are
inteqnjependent with the child's socio-cultural environment
whicrl is transmitted to him throuch the lincuistic process.

The code the child brinms to the school

symbolizes his social identity. It re-

lates him to his kin and to his local

social relations. The code progressively

orients the child towards a pattern of

relationships which constitute for him his

psychological reality, and this reality is

reinforced every time he Speaks. (p. 259)

An example of the reinforced psycholocical reality

‘Whictl Bernstein provides, is the difference between the
following two statements:

"I'd rather you made less noise, darlinn."

"Shut up."

The liist statement is meaningful to a middle class child, but

111

what is important to stress is the fact that the middle
class child has learned to be able to reSpond to both state-
ments, and both are differentially discriminated within a
finely articulated world of meaning. The workinp' class child
only has meaning: for the one and if he hears the other he
will translate it into his own language as "Shut up!" Thus
the elaborated lannuaqe code is mediated throurrh the restricted
code of the working~ class child in order for the message to
be personally meaninr'ful to him. Where the working class
child cannot make this translation, he fails to understand
and is left puzzled, according: to Bernstein.

The middle class child's relation to the environment
"is such that his range and expression of discriminatino'
verbal responses is fostered by the social structure from
the beginning. A virtuous circle is set up which is con-
tinually reinforced, for the mother will elaborate and
eXpand the embryo personal qualificatory statements that
the child makes." (Bernstein, 1958, p. 163) Thus the
greater the differentiation of the child's eXperience the
greater will his ability be to differentiate and elaborate
ObJects in his environment. The point thus is not that the
middle class child has more experiences but that they are
more linguistically differentiated for him, thus in effect,
creating; more experiences.

Bernstein provides the following example of the

difference between working class and middle class mother-

 

 

a r
b .. uq . . .
.v o s .5 O O ,
Au 5 n . M“ ~ .» c 5. .. E s .
. . 5 of , . I. l 5 LL 1; as
a. . a: .r v on _ h... n‘. by r . . fly a A \II.‘
A» C» Dc 9“ .n.. h. w .1. 0.. m“ .0... A.» a
fizv 2‘ A C ~7— . 9w figs “- Lb A V.” C
. at. a». v.“ u A). “m. r. +v rc
0‘ sl 6 C .M. 1 his 5 Y\.
C. e c «m» .th
,\

 

 

 

Q .
his Md. .s
«T. n » Av
9 AC
ru. .NH Nu
m . H4 3
a: .MVL

15

child interactions. The pairs are on a bus and the child
is setting on the mother's lap.
Restricted Code Users:
Mother: Hold on tight.
Child: Why?
Mother: Hold on tight.
Child: Why?
Mother: You'll fall.
Child: Why?
Mother: I told you to hold on tight didn't I?
Elaborated Code Users:
Mother: Hold on tight, darling.
Child: thy?
Mother: If you don't you will be thrown forward
and you'll fall.
Child: thy?
Mother: Because if the bus suddenly stops you'll
jerk forward on the seat in front.
Child: Why?
Mother: Now darling, hold on tightly and don't
make such a fuss.
Although the example is very concrete and short,
it frithiqhts a couple of issues. A whole ranqe of potential
leaiuming and connections have been cut out of the first
example»by the use of the cateaoric statement. Natural
(nxriosity is blunted. The notion of authority and status,
INJtential social power, is revealed quickly. In the second
examplethe child is eXposed to a number of lincuistic
Connections, reasons and sequence. If challenged a new
SGt cxf reasons is employed. Although social power is finally
reveeuled, it is much later and under different conditions.
The lireouent use of the categoric statement limits learning

and (Hiriosity and induces a sensitivity towards a particular

tyne of authority in which social power is quickly revealed-

16

The categoric statement becomes a part of a language which
narrows the range of stimuli to which the child responds.

Pestricted and elaborated codes may thus be considered
as 'programs' for the human information processor and,
therefore, as with the machine language of a computer system,
different kinds of information are perceived and accepted;
the information may be interpreted differently; and thus
behavior which may result from the 'same input' may ouite
possibly differ.

Because linguistic codes determine the 'planniny pro-
cedure' or 'proqram' for information processino', and because
an elaborated code produces a planninp‘ procedure which
promotes a relatively higher level of syntactic organization
and lexical selection than does a restricted code, that which
is made available for learning- by the two codes is different.
Learning in this context, refers to what is significant, to
what objects are made relevant, social, intellectual, and
emotional objects.

Bernstein's theory may sound very similar to that of
the Sapir-Whorf tradition, however its basic assumption is
the inverse. For Whorf (1956), the language system is the
b51310 and primary force creatinrz different social and cultural
environments, thus creatinp: different ways of viewing and
int€31"oret:in9; the world. For Bernstein, elements of the

SOCHal-cultural structure are basic and underlie the formation

9 i .
n

no. A
AU no

4.
u

I” f":

-\-

c—xc'rar.~
facili

A

estigv

fin
.14

17

of a particular form of linguistic code, and the code rein-
forces the socio-cultural environment, or ways of viewing
and interpretina the world.

In Bernstein's 1959 Daper he discusses the important

functicwivmdch language plays in influencing behavior:

Lanquaae is considered one of the most
important means of initiating, synthesizing
and reinforcinw ways of thinkinm, feeling

and behaving which are functionally related

to the social group. It does not of itself
prevent the expression of specific ideas or
confine the individual to a given level of
conceptualization, but certain ideas and
generalizations are facilitated rather than
others, that is, the language used facilitates
development in a particular direction rather
than inhibiting all other possible directions.

(0. 312)

 

A restricted code facilitates the construction and
excrmnage of communalized symbols and an elaborated code
facilitates the same for individualized or personal symbols.
Regulation and control patterns inherent in an elaborated
code, irfluces its users to implications of separateness
aIKi<iifferences in phenomena and points to the possibilities
Ifllr the organization of eXperience inherent in a complex
Conceptual hierarchy. These two codes are induced by the
SOCiitl relationships or control patterns within the family.
The 1Jitellectual and social procedures by which individuals
PGIaIHe themselves to their environment may very much be a
QUGStxion of their 'Speech models' within the family and the

COdes; these speech models employ. Middle class and associated

l8

strata socialize their children to possess both an elaborated
and restricted code; while children socialized in some
sectixyns of the lower working class strata, can be expected
to pcnssess only a restricted code. To succeed in school it
is critical for the child to possess an elaborated code, or
at leuast be oriented towards it.

The ability to switch codes for various social rela-
tionsfliips, controls the ability to switch roles. Thus one
of the consequences of these codes is the ability to empafifize
or svxitch roles--a necessary and important feature of effective
comnnuaication (Berlo, 1958).

Althoumh both Whorf and Bernstein posit different
prinwrry functions of human behavior, the lanmuane system vs.
true social-cultural system, respectively, both would allow
ftn? an interdependence or interacting influence of the
lineuistic code and the social structure in which the code is
used.

Watzlawick et a1. (1967) in discussinm the influence
Of conmmnicative behavior, provide a similar view as the
sociolinquistic theory of elaborated and restricted codes
Fumesented by Bernstein. The authors sumwest two functions
or a.rnessa2e, a 'report' and a 'command' function. A report
functxion is basically the information which is transmitted
bEtMNEen two individuals and the command function provides

a definition of the relationship between the two individuals.

. . 1a ‘ . . i 11 . . o 0 . Vlu — o f,. 5|... Cs 1— I—
... it 3 .. _ .l "A E . a o t ~ .1 n” 3 m. 9 .C Y .3. Q» L ..
s S n A. . .3 a S n. h l a o n. e C E w e .3 .3. t .2
n . C 3 E a. . T A; I E .1 .l e o. .2 S a E .1 F E .1 h.
C 3 w .3 C C C E P a l n. .C t U C o i v v a T. .w. 2“
r I'll

 

 

19

In view of Bernstein's conception, different sets (mroups)
of individuals appear to employ characteristics of a lancruacre
code which not only restrict the information value of the
message transmitted but also frequently initiate a different
definition of relationship. In essence it is the exnlicitness
of intent and personalization of relationships which are
provided by linmuistic structures and lexicon in an elaborated
code, but the same significances are more frequently and
explicitly transmitted nonverbally and by context for a
restricted code-user. Nonverbal information as well as
particular lexical and grammatical alternatives (para-
linquistic features of lanauao'e) may provide 'command' or
relationshp functions. Thus for elaborated code-users,
both 'structure' and 'content' are predominantly communicated
through the linguistic code, while for restricted code-
users, many of the 'structural' relationships are communi-
cated nonverbally or in codes other than the linrruistic one.
anverbal eXpression such as gestures, paralinguistic
features, facial eXpressions, bodily movements, are termed
'immediate or direct expression', while words are termed
'mediate or indirect expression'. The importance of direct
and. indirect expressions is the relative emphasis and use
made of one or the other. The working: class child's languame
With a hisrh proportion of commands, short simple statements

and questions with simple, descriptive, tangible symbolism

20

of a low-order generality, relies heavily upon nonverbal means
of expressing; personal feelings. The lanmuap'e does not have
the rich vocabulary and grammatical complexity to differen-
tiate and finely distinguish personal qualification as does
the middle class child's language. This form of language

has many implications for the structuring of experiences and
relationships with obiects.

Although the middle class languaqe code—user can use
many of the nonverbal means of expressing“ feelings, they
take second place to his sensitivity to verbal differentiations.
The middle class child at an early acre becomes sensitive
to a form of languao'e use which is relatively complex and
which in turn acts as a dynamic framework upon his perception
of objects. There is a pressure in the middle class social
structure to intensify and verbalize an awareness of
separateness and difference which increases the significance
Of objects in the environment.

Thus the distinction between structure and content
is one of degrees within a conceptual hierarchy. What is
implied is this: "Where there is sensitivity of content only
the simplest logical implications or boundaries of the
Structure will be cognized." More specifically, certain
aspects of an object will not register as meaningful cues;
01" if they do, the verbal response will be inadequately

dEteI‘mined .

21

Because of the working class sensitivity to the con-
tent of’objects and to personal qualification by means other
than language structure, workina class members learn to
respcnud and make responses to cues which are immediately
relevarfln Stress on the present in the mean§_of communi-
caticni preclude the understandina of the meaninafulness of a
time <:ontinuum other than of a limited order. Time—span of
antixxipation is brief and the workinv out of connections
betweemlneans and distant ends is very immature. An important
conseuquence of this patterninm is that it produces a descrip-
tive <3ognitive process, e.g., crude loqical connections be-
tweeui events A, B, C, and D, and that sustained curiosity
is TMDt fostered or rewarded as answers to questions rarely
learl‘beyond the object or further than a simple statement
atmnxt the object (Schatzman and Strauss, 1955). The social
strnuiture continues to reinforce the early patterning of
perception.

We have stated above that Pernstein talks about
larumiage codes (elaborated and restricted) or styles of
lanauaee behavior in terms of a functional relationship with
societl interaction. Particular family control patterns or
SOCiéfil structures influence the lineuistic code learned and
the ruature and function of the code in human interaction or
cOmmunicative environments, i.e., the lancuace code creates
Dartixzular cognitive styles of behavinv. At this point let
us ttxrn to the underlyinr cultural and social subdeties which

Rive :rise to styles of interaction throuqh linguistic code media.

22

Bernstein has found that social class is a variable
which provides a sross distinction between restricted and
elaborated linmuistic codes. Social class variables which
have been employed by Bernstein and others in Fncland as
well as by American researchers are those of occupation,
income, mother's education, and place of residence. Fernstein
(1962a) also employed the variable of 'which school the
younqster attended' a repular high school or a vocational
technical school. However, in his recent writings (1969)
he placed greater emphasis upon the home environment and
patterns of interaction or control within the home, specifi-
cally between the mother and child--the mother beinq the
dominant adult lanvuaqe model during the early years of the
child's laneuage develOpment.

Bernstein's arcument seems to suggest that linouistic
differences in education as well as other test measures are
results of a different code; and a different code is a result
of different family styles; and family styles may be cenerally
distinguished by the notion of social class: lower-working
class and middle class.

A dynamic interaction is set up in the middle class
family and social structure. fihe pressure to verbalize
feelings in a personally qualified way, the implications of
the language learned, combine to decide the nature of the
cues to which he responds--structural cues. And an orien—

tation towards structure, rather than towards content as

 

w . W. Le T. o‘ _ w. 4 (A: Y“ m A_ .- ‘ D . . 3. CC. V 31.. ‘ u o. as \ 5
S 3 S n. E r S .. . .3 2. .F. a. t l S t o .. i. a. B
e a c I .1 a. u .. i H C 2 ~ , e .C n. 9. a: S a . .3 3 C S
C .l u ,7. Z .1 a. 2.. i n. .L .2. E a C P O .L 3 .J M. .1
.3 c... .2 3 .. _ S S a. . u G t ”a C C r r. n. .3 ._ .. e u.

 

 

23

does the working class social structure, allows a number of
alternative interpretations and meaninms of an object which
thus increases the area and intensity of the child's curiosity
and receptiveness.

The middle class child crows up in a social environment
where spatial and temporal seouencinm is finely distincuished.
Thus the child is inculcated with a fine sense of relation-
ships between future aoals with present activity. The
sensitivity to the structural relationships among objects
facilitates the development of conceptualizations which are
unavailable to working class children.

Working class family structure is less formally
organized than middle class family structure in relation to
the development of the child. Althouqh authority is Specific
within the family the values expressed do not provide the
carefully ordered universe-—spatially and temporally--
of the child. Uhus the specific character of long-term
goals is replaced by more general notions of the future.
Present activities have a much greater value and therefore
postponement of present mratifications for future ones is
difficult.

The language between mother and child is restricted
for the workinm class child: one which contains few personal
qualifications, for it is a lanouaae where the stress is on
emotive terms employing concrete, descriptive, tanvible and

visual symbolism. The nature of the lancuame tends to limit

2M

verbal expression of feeling. Feeling between mother and
child tends to be communicated nonverballv. Thus feelings
are less differentiated by the workinv class child throunh
the medium of languame. Cues responded to in his environ-
ment are primarily of a qualitatively different order. He

is sensitive to the content of objects. Of critical im—
portance is the type of lancuame-use upon which value is
placed, for once a value is so placed, then that lanruame-use
will reinforce the emotional disposition which resulted in
the initial preference.

Bernstein (1959) has distinpuished two forms of
linpuistic codes which result from the kinds of family and
social environments described above. Fe succests that a
family which employs a restricted lanpuame code provides a
child with a languame environment characterized by the
followinm:

1. Short, qrammatically simple, often unfinished

sentences with a poor syntactical form stressin'T
the active voice.

2. Simple and repetitive use of conjunctions
(so, then, because).

3. Little use of subordinate clauses to break
down the initial catecories of the dominant
subject.

H. Inability to hold a formal subject throuph

a Speech sequence; thus a dislocated infor-
mational content is facilitated.

5. Rigid and limited use of adjectives and adverbs.

6. Constraint on the self-reference pronoun;
frequent use of personal pronouns.

10.

11.

On

25

Frequent use of statements where the reason
and conclusion are confounded to produce a
cateaoric statement.

A larme number of statements/phrases which
signal a requirement for the previous speech
sequence to be reinforced: 'Wouldn't it?',
'Vou see.', 'Vou know.', etc. This process is
termed 'sympathetic circularity.’

Individual selection from a proup of idiomatic
phrases or sequences will freouently occur.

The individual qualification is implicit in
the sentence ormanization; it is a languame
of implicit meaninm.

Symbolism is of a low order of aenerality.

the other hand, a family which employs an elaborated

linguistic code orients the child to a lancuame environment

characterized by the following:

1.

2.

Accurate prammatical order and syntax revulate
what is said.

Loaical modifications and stress are mediated
through a grammatically complex sentence

construction, especially throuvh the use of a
range of conjunctions and subordinate clauses.

Freouent use of prepositions which indicate
loqical relationships as well as prepositions
which indicate temporal and spatial contiquity.

Frequent use of the personal pronoun 'I', and
impersonal pronouns, 'it', and 'one'.

A discriminative selection from a range of
adjectives and adverbs.

Individual qualification is verbally mediated
through the structure and relationships within
and between sentences.

Expressive symbolism discriminates between meanings
within speech sequences rather than reinforcing
dominant words or phrases, or accompanying the
sequence in a diffuse, generalized manner.

26

8. A lanruame use which points to the possibilities
inherent in a complex conceptual hierarchy for
the orranizinr of eXperience.

Behavioral scientists in the United States seem to
support Bernstein's conception of the home as the dominant
influence in determining the kind of linguistic code learned
and used by children. Bloom (1965) has produced 10“ paces
of annotated abstracts from empirical research papers con-
cerned with lanmuaqe and academic achievements. He writes
on the basis of those documents: "...that the home is the
single most important influence on the intellectual and
emotional development of children particularly in the pre-
school years." (p. 69) Pe continues by sayinm that for
adequate language develOpment, there is a necessity of
frequent interaction between child and parent at mealtime,
playtime, and throuqhout the day. Central to general learnina
and language development is the amount of interaction
between parents and their children, and the amount of
corrective feedback.

The notion of the importance of the expansion of
language by the mother is also part of Power Brown and
associates' (1965) theory of language development of the
child. Providinm seouencinm and sets of reasons for the child's
'Why?', 'How come?', etc. questions, are vital to the child's
growth. As is the immitation with slivht modification which
provides 'correct' grammar plus many of the function words

and auxiliary verbs, etc., which the early child leaves out.

27

Dave (1963) found that the home has the greatest
influence on the lanruaqe development of the child and the
least on the skills tauaht primarily in the school. Fe
further concludes that it is not so much the status
characteristics of parents but what the parents do in the
home in terms of interaction with each other and their
children, which are most influential on the achievement of

their children.

LaboV's Theory

 

Althoumh Bernstein, Floom, Brown and Dave all suvrest
that the home is the most siqnificant influence upon the
child's language development, another sociolinpuist in the
United States, Labov, makes a stronm argument for the peer
group emerainm as the most simnificant influence upon the
child's languame behavior. Labov (1970) states that in the
sociolinguistic study of lanvuaqe learninp, we can bemin
with the fundamental observation that "children do not speak
like their parents." (p. 33) At first plance, the statement
seems somewhat surprisinm as we obviously learn to Speak
from our parents. If the child's parents speak anlish, and
he qrows up in the United States, he will certainly have
anlish as his native lancuape. "Vet in almost every detail,"
states Labov, "his Pnclish will resemble that of his peers
rather than that of his parents." (p. 33) Labov does

aaree that linguistic codes may differ between workinm and

28

middle classes. However, he states that children across
both class levels will have a more homogeneous code than
that code of parents and children within either of the

two classes. The present study will provide a more critical
test of Labov's hypothesis that peers rather than parents
provide the most significant influence upon the child's use
of a linguistic code.

Although Labov admits that as yet we have no thorough-
going studies of the relation of parent, child, and peer
group, all of the available evidence shows that the peer
group has the greatest influence. Labov continues:

Most parents are not aware of how system-

atically their children's speech differs

from their own; if they do inquire, they will

be surprised to find that there is no fixed

relation between their own rules and those of

their children. Instead, it is the local

group of their children's peers which de-

termines this generation's speech pattern.

This is the case with rules of non-

standard urban dialects as well as the more

neutral rules of regional dialects con-

sidered here. (1970, p. 34)
According to Labov, the full force of the peer group doesn't
appear in the six-year-old in the first grade, but does
occur in the fourth and fifth grade, when the ten—year-old
begins to come under the full influence of the pre—
adolescent peer group. Dialect differences and sharp
downward trends in school records appear at this time and
seem to not be "unconnected with the fact that peer groups

present a more solid resistance to the schoolroom culture

than any individual child can." (1970, p. 34)

29

Labov, then, as with Bernstein, has based his
sociolinguistic theory of lincuistic codes unon indirect
research and unon conceptualizations of social nsycholocy,
education and sociology. The inconsistency which seems
apparent has not been critically tested. The present study
will provide a test of whether or not it is the narent or
the child's beer proun that has the greatest influence in
determining the child's speech patterns. Limitations of
the present research are that it is a correlational study and
is conducted in the United States. Also the researcher has
made no attempt to select a random samnle from the pooulation

of teenagers or the population of all children.
Summary

fernstein has discovered systematic differences be-
tween children of 6 years of age as well as teenagers no
to 15 years of age. From these differences, one can make
some assumptions about the nature of family communication
styles or control patterns. If similar differences in
language use can be found among American families, then it
would be important to study family interaction patterns over
time to obtain further insight into relationshins among
language development and academic potential in children.
The present study is concerned with discovering language
differences amonv heterogeneous families as rewards social

class variables of education level of mother and income level

30

of the family. In addition simple observation of physical
characteristics of the home, both external and internal was
made. Analyses of lincuistic codes of teenacers, their
mothers and their peer Groups will be conducted.

If sicnificant variables are found which distinguish
elaborated and restricted codes, then implications for
future research as well as practical application in the
formal education system appear necessary and beneficial. It
is important to point out that researchers in Pritain have
found that the use of a restricted or an elaborated code is
relatively independent of intellicence and other psycho-
lOpical personality characteristics. Ibwever, such
variables may influence the level at which one operates
within a particular code.

Whe research proposal briefly described above (that is,
determining whether or not similar general linguistic codes
are used in 'white' American subcultures as has been found
in Britain) seems to be a necessary pre-reouisite to further
analyses and messave manipulations of lincuistic and non-
linauistic codes as they apply to formal education. fihe
relationships amonc social structure, lanmuage-use, and
subsequent behavior have direct bearinc on questions such
as the followine: (1) that is available in the environment
to be learned?; (2) that are the conditions for learninc?;

(3) Vhat are the constraints on subseouent learninc?;

31

(D) What are the major reinforcinc processes?; and (S) that
are the significant linguistic variables which facilitate
intellectual and cognitive crowth, or facilitate human
functioning and adaptinc in a heteroceneous society.

then children respond differently or rive 'wronm'
answers, it is not so often that they are wrong as that they
are answering another ouestion. The two codes described above
produce different restrictions on what is perceived as
relevant and the interpretations of that which is perceived.
An explanation of 'wrona' answers mivht partially be given
by the lincuistic characteristics of the two codes.

In the next chapter specific research related to
elaborated and restricted codes is reviewed. Specific

research questions and hypotheses are also presented.

CHAPTFQ II

PFSFAWCH LITERATURE AND HVPOTFESFS

Presummary

 

In the previous chapter we reviewed the basic con-
ceptual and theoretical sociolincuistic literature relevant
to the emersence of two linmuistic codes--elaborated and
restricted. It was sur"ested that the two codes emerme
from a difference in social class level of families with
particular patterns of social interaction. One sociolincuist,
Labov, presents an alternative conceptual framework from
that of Pernstein, in that he hypothesizes that the child's
peer groups have a greater influence upon the child's lin-
guistic code than do the child's parents. pesearch reported
in this study will attempt to provide empirical data relevant
to the apparent conceptual differences between Fernstein
and Labov.

The present chapter reviews empirical research
relevant to the emercence of elaborated and restricted codes.
In addition, general questions and research hypotheses based
upon theory and research relevant to the sociolinguistic

study of elaborated and restricted codes are presented.

32

33

Fxperimental Desearch

 

Three major exnerimental studies have been conducted
which provide information directly relevant to the socio-
linguistic theory eXplicated by Professor Fernstein. The
authors of the research, Bernstein, Lawton and Pobinson,
are presently associated with the Sociological Pesearch Unit
at the University of London. Fach of the studies deserves

discussion in depth.

Fernstein's Qesearch

 

Two major experimental papers by Professor Bernstein
are currently available to the experimenter. (Bernstein 1962a,
1962b) The same subjects were used for the two separate
analyses desiened (l) to see whether the two hypothesized
codes - restricted and elaborated - were associated with
social class, (2) to see whether the orientation to one or
the other codes was independent of 1.0., and (3) to see what
linouistic units were critical in distinguishing the two
codes.

The samples of subjects consisted of sixty-one boys
aged fifteen to eiphteen of workinm-class backcround
(messenger boys -- none of whom had been to mrammar school)
and forty-five boys of the same age from a public school. A
tape recorded, relatively undirected discussion was taken
with all Proups on the topic of the abolition of capital
punishment. A sample of 1800 words which followed the first

five minutes of the discussions was used for the analyses.

Study I

In Bernstein's first analysis, he found that overall

social class differences were sipnificant as predicted:

l - Working-class (WC) subjects used a lonmer
mean phrase lenoth, Spent less time pausina
and used a shorter word lenqth.

2 - Foldinm nonverbal intellimence (Paven Droo'ressive
Vatrices Test) constant, social class differences
were found in the same direction.

3 — Foldins verbal (”ill Pill Vocabulary Side Form I
Senior) and nonverbal intellidence constant,
social class differences were acain found in
the same direction except for word length.

H - Within the middle-class (MC) croup the sub—
aroup with superior verbal intelligence used
a longer phrase lencth, a faster rate of
articulation and a lonper word length.

5 - Within the HC group the sub-croup with the

averame 1.“. profile Spent less time pausing.

Bernstein concludes that there is a difference between
workinq-class and middle-class linguistic codes which are
revealed in verbal planninm procedures as measured by
hesitation phenomena. Although I.“. does make a difference
within a social class group, it appears to be independent

of whether a 'restricted' or 'elaborated' code is employed.

Study II

35

Eernstein's second published paper "Social Class,

Linguistic

samples as

Codes and Grammatical Plements" employed the same

described earlier with analyses made on only ten

middle-class subjects and fourteen workinc-class boys.

Results of
as follows:

1 -

Eernstein's analyses of crammatical elements were

Pb social class differences were found in the
proportions used of finite verbs, nouns,
prepositions, conjunctions and adverbs.

"I think" (referred to as emo-centric sequence)
was found to be much more frequently used by

MC boys, whereas the sympathetic circularity
sequence (e.p., 'wouldn't it', 'isn't it',

etc., referred to a socio—centric sequence)
occurred more frequently in working-class speech.
The MC croups used a higher proportion of:

(a) subordinate clauses; (b) complex verbal
stems; (c) passive verbs; (d) total adjectives
and uncommon adjectives; (e) uncommon adverbs
and conjunctions; (f) 'of' as a proportion of
'of', 'in' and 'into'; (a) personal pronoun 'I'.
The WC groups used a himher prOportion of (a)
total personal pronouns, and (b) 'you' and

'they'.

36

Once acain Pernstein defines the two codes in terms
of predictability. The structural elements are hichly pre-
dictable in the case of a restricted code and much less so
in the case of an elaborated code. Althourh some of the
criteria (e.r., uncommon adjectives) employed for
differentiatinc the two codes are quite arbitrary, he has
provided data in the form of linruistic units which support I

differences found in hesitation phenomena, verbal planning,

E .flfizxfi. “4—3
a.

and abstract-concrete levels of lanauave.

Bernstein suggests that the best sinmle indicator of
the two codes is the prOportion of subordinations to finite
verbs. He restates that his findinss clearly indicate that
speech orientation to the two codes and verbal planninq
processes which they entail, are independent of measured in-
telligence.

Bernstein's research was stimulated by three questions.
Are the elaborated and restricted codes associated with
social class? The answer is 'yes'. Is an orientation to
one or the other code independent of I.0.? Again Bernstein
found the empirical answer to be 'yes'. The third question
concerns the linmuistic units which are critical in dis-
tinguishing the two codes. Significant linguistic variables
are summarized in the list of four conclusions stated above.

At this point we can say that the sociolinguistic
theory of Bernstein discussed in Chapter I is generally

supported. We turn now to research conducted by Denis Lawton.

"

 

37

Lawton's Pesearch

Lawton (1968) conducted research for three reasons:

1 - to provide data which would support or refute

Bernstein's research and theory°
3

2 - to extend the range of evidence by collectinc

3 - to investipate other kinds of speech situations

'discussion' speech from some boys youncer than

the subjects used by Bernstein; and

F -‘t.._')' -‘.‘K
.0 ._
r

and the written lanpuace of the same boys.

Fecause of the larce amount of work entailed in

linguistic

analyses, Lawton limited his sample to four Groups

of five bovs each: five fifteen vear-olds from each of the
z, .9 Q.

two social

classes and five twelve year-olds from each of

the two classes.

Each boy wrote four essays with a 30-minute time limit.

The topics

1 _

were:

A story ('Urashima the fisherboy') was read

to the croup and they were immediately asked
to rewrite it.

'Pome'.

'My Life in Ten Vear's Time'.

To explain the name of 'soccer' to a foreicner

who has neither heard of it nor seen it played.

In addition, each boy was to complete two Sentence Completion

Tests.

38

The second part of the study was to attempt to
replicate Fernstein's informal discussions by having the boys
participate in a thirty-minute discussion of capital
punishment.

The third part of Lawton's research consisted of
individual interviews with each boy. The purpose of the
interviews was to provide a context wherein it was possible
to control the level of coding difficulty, and to contrast
the performance of the groups differing in age and social
class in two kinds of speech situations labelled 'Description'
and 'Abstraction' (Goldman-Eisler, 1961).

The interview was divided into three sections:

1 - Four sets of picture cards: each set consisted

of four or five cards which conveyed a story.
The subjects were first asked to describe the
picture (description), and then to abstract
the point of the story (abstraction).

2 - Questions about school involving first des-

criptive, then abstract language.

3 - Four questions involving moral judgment

(abstraction).
Results of Lawton's research suggest the following:
A. Written Language
1 - Clear evidence was found that the working-
class/middle-class difference in usage of
Restricted and Elaborated Code applies to

written work as well as Speech.

2 -

39

Although there were social class differences
even on narrative-description essays, the
social class differences were greatest on
essay subjects which enabled abstract
writing to be selected.

A new Sentence Completion Test (details not
given) was devised which was successful in
distinguishing working-class and middle-

class boys in their use of subordinations.

Group Discussion

In the discussion situation, planned as a

replication of the Bernstein eXperiment,

linguistic results remarkably similar to

Bernstein's were found, not only in the direction

predicted, but in the actual numerical scores.

Individual Interviews

1...

Social class differences were still in
evidence, but

the actual numerical results were quite
different from the figures relating to
'discussion speech'. Thus Speech was

found to vary not only according to social
class, but also according to general context

of situation, i.e., the mode of communication.

HO

3 - Interestinq differences were found in
relation to description and abstract
languaqe, which succests that workinm-class
boys can be made to use something which is
at least approaching an Flaborated Code.

All three studies of Lawton show significant differences
in use of lincuistic forms by the two social classes. These
differences are greater at awe fifteen than age twelve. The
differences are stable across various modes of communication
as well as subject matter and are less distinct when subjects
are forced to speak or write about abstract subjects.

Once again the theory posited by Bernstein is basically
substantiated. Support for the theory is further expanded
to include various modes (communication mediums) of inter-
actions as well as an extension of the age level of boys
from 15 years to 12 years. Results of Lawton's study
also suggest that the elaborated - restricted codes are on
a continuum and that under appropriate conditions, a re-
stricted code-user may be able to employ some of the
characteristics which an elaborated code-user employs.

The final research reviewed in this chapter is that

of Professor Robinson.

Pobinson's Pesearch

Pobinson Questions the validity of Bernstein's argument

that many workinn—class children, especially those from the

41

lower working-class, have no access to an elaborated code.

Both Bernstein (1962a, 1962b) and Lawton (1968) had reported
social class differences in language consistent with Bernstein's
sociolinguistic theory.

Four asoects of the experimental desiwn employed by
Bernstein and Lawton are questioned.

l - The tonic of communication - Capital Punishment. p:
Pobinson argues that there is a possibility of differential I
amounts of relevant information available to both groups. L~

2 - The type of receiver. The fact that working-
class boys were talking with working-class boys who might
expect "restricted" code utterances, may have deterred
them from using the 'elaborated' code.

3 - The structure of the communication system. Because
of the unstructured conversational atmosphere, the working-
class boys may have been encouraged to use a 'restricted'
code.

H - The mode of communication. Perhaps speaking in
a group did not offer the best opportunity for working-class
boys to display their 'elaborated' code. Writing might allow
use of an 'elaborated'code.

Robinson used over 120 boys and girls with aces
between twelve and thirteen. All subjects wrote two letters --
cine 'informal' and the other 'formal'. The informal letter

Wzas to be written to a friend whom the subjects hadn't seen

U2

for a long time. Each subject was to write about the latest
news and gossip. The 'formal' letter was addressed to the
'Governor' of the school explaining reasons for wanting to
go to a particular location.

On a large number of linguistic variables Bobinson
found that on the 'formal'letters there were only a few
statistically significant differences employing Wilcoxon
Matched-Pairs Signed-Banks Tests. The informal letters
exhibited more significant differences.

Robinson concludes by sayinq that the restricted
number of variables examined by Bernstein and Lawton, and
the possible instability of differences found in his own
study stemming from small language samples, make general
evaluative comments tentative and imprecise. Lexical
differences seem to support Bernstein and Lawton; however,
structural results are not so consistent.

Robinson (1965) conducted one other analysis upon
his subjects. He employed a special use of Cloze procedure
(Taylor, 1953) to determine whether or not working class
children would use a more predictable code (structural and
lexical alternatives) than middle class children.

The subjects were given thirty-one experimenter
created sentences, each sentence on a different page. One
word was deleted from each sentence. Each subject was to

fill in the first word he thought of, then put in another

» I

43

word below it, and finally put down as many other words
that he could think of which might 'fit', the deleted word.
At a later time subjects ranked their resoonses havinm the
Opportunity to change their first response.

Robinson deleted the following words; M nouns;
u adjectives; M verbs; U prepositions; 2 conjunctions; and
a number of others, unspecified. Results indicated that
middle class boys used a wider rance of words and working

class boys had greater conformity of responses. General

”:2;
..
.

results supported Bernstein's conceptions of differential
predictability within middle and working class codes.
Robinson summarizes with the statement that "Cloze procedure
seems to be a useful way of eXploring the nature of languaae
differences." (p. 55)

A number of problems arise in attemptinq to compare
the works of Bernstein—Lawton and that of Robinson. The
following speculations may be operating:

1 - Writing a letter to the 'Governor' may have
decreased the use of an elaborated code for
middle-class children because of the uniqueness
of the task.

2 - Different research designs were employed.

3 — Boys and girls were used by Robinson and only

boys were subjects for Bernstein and Lawton.

an

H - Different topic areas and constraints were
employed.

5 - Lincuistic units studied varied in many
instances.

6 - Awe differences of the children varied.

Robinson's research menerally supports that of

o 7—1 I 1 o o ‘
Bernstein and Lawton. vurther eVioence lS prov1ded which mz‘
suggest that a restricted code-user may be forced to use a a
code more similar to that of the elaborated code-user.

As the present study did not analyze 'writinw' of the
various subjects, it cannot examine Robinson's consideration
that perhaps workind-class children have an elaborated code
available but are not stimulated to use it. Dobinson's
alternate summestion is that linmuistic differences between
social classes may be operapt at early aces, diminish around
age 10 or 12 because of common school influences and then
widen again around ace 15. The 'ame hypothesis' was not
explored in the present research. Only one communication
mode is employed in the present research - a family discussion.
'Thus the possibility that teenavers who predominantly use a
restricted code could be forced to employ a code more similar
to an elaborated code is not tested.

The followinm section discusses peneral research
questions, a summary of hypotheses derived from the reviews
in Chapter I and II, and a rationale for each specific hy-

pothesis.

“5

Research Areas and Hypotheses

 

Research Areas

A review of the sociolinguistic theory posited by
Professors Bernstein and Labov and an analysis of three
major research studies suggests the following two relation-
ships:

1. Individuals identified as belonging to a lower
socio-economic class will employ a 'restricted' language code E
whereas individuals representing a middle class environment I
will employ an 'elaborated' communication code. The two
codes demand particular information processing programs
which are ouite different in terms of orientations toward
other people, objects and ideas as well as different levels
of verbalizing eXplicitness or intent of those objects.

The two codes are differentiated in terms of the
probability of predicting which structural and lexical
elements will be used to organize meaning. An elaborated
code-user (middle class) will select from a relatively ex-
tensive range of alternatives. mhe restricted code-user
(lower working class) will employ a limited number of the
linguistic structures and lexicon.

2. Social class has a greater influence upon the
linguistic code used by individuals than peer wroup influence.
Specifically, Bernstein, though not Labov of the United
States, predicts that children will employ a communication

code more homOgeneous with other members of their family

U6

than with their peers representing a different social class.

In the present study, a teenage boy will use a code more

like his mother than like a teenager from another social

class level. Similarly, a mother in a family will employ

a code more similar to her son than a code used by the maternal

parent from a different social class.

i

The two general areas described above lead to the

specific hypotheses stated below. The hypotheses are designed

F vmflifr-m m'
A __
A)

to test whether or not linguistic differences relevant to
elaborated and restricted codes would be found among a
heterogeneous group of 16 families. Descriptive data obtained
through content analyses of the Speech of the family members

are used to test the hypotheses.

General Hypotheses

Ten measures of a communication code will be employed
to determine if significant differences are revealed between
two samples of families - one designated 'working class' and
the other 'middle class'. On all ten linguistic variables
the following two general hypotheses are suggested on the
basis of sociolinguistic theory and research related to the
emergence of elaborated and restricted code systems.

1 - Members of families of the 'working class'

sample will exhibit significantly different
scores on all dependent linguistic variables
than members of families of the 'middle class'

sample.

“7

2 - Mothers and sons of each sample will exhibit a
more homOgeneous use of linguistic variables than

will mothers or sons across samples.

Specific One-Alternative Hypotheses

H1: Working class families will exhibit a

greater prOportion of pronouns/total words

than will middle class families. “H
Numerous studies (Lawton, 1968, Chapter III) regard H

a diminishing percentage of pronouns as an indication of
increasing maturity of writing, abstract writing and more
impersonal writing. Use of personal pronouns also suggests
an insufficient Specificity whereas a noun or noun phrase
would be more appronriate and indicate a greater Specificity.
One may also interpret the use of a pronoun combined with a
non-verbal signal as adeouately communicating intent and
Specificity in many concrete situations.

H2: Working class families will exhibit a

greater prOportion of pronouns 'you'
and 'they' to total pronouns than will
middle class families.

A ready identification or implicit agreement in
referent is implied by the use of 'you' and 'they'. Inasmuch
as referents are not finely differentiated, the global term
'they' will be adapted as a general label. Both pronouns are

nonSpecific and arise out of an attempt to concretize eXperiences.‘

U8

.4.

H3: Working class families will exhibit a
greater proportion of socio-centric seouences/
total words than will middle class families.
When used repeatedly, socio-centric seouences dis-
courage further analysis, and close the discussion at that
particular level. Curiosity is limited. Veanings signaled
in this code tend to be implicit and so condensed. The

socio~centric sequences may be transmitted as a response

of the speaker to the condensation of his own meanings. The

l inn—mu;
F: - .1

speaker reouires assurance that the message has been received.

HM: Linguistic codes of working class families
will exhibit a higher Cloze score than will
the linguistic code of middle class families.
Cloze Procedure is a measure of readability or com-
plexity of a message. It also may be a score correlated
with an information theory measure of redundancy. According
to Bernstein, a restricted code uses fewer grammatical and
lexical Options than an elaborated code. Thus predictability
should be greater for the restricted code and the Cloze score
would thus be greater than for an elaborated code.
H5: Middle class families will exhibit a
higher prOportion of the pronoun 'I'/total
pronouns than will working class families.
She pronoun 'I' increases the differentiation of self.
A restricted code does not encourage Specificity or separa-

tion of self from others who are considered as a common or

H9

like group. A middle class family stresses the person-
relationships and thus qualifies and differentiates what he
would say from that of someone else. Thus a middle class
member uses the pronoun 'I' more frequently than a working
class member.
H6: Middle class families will exhibit a
higher probortion of ego-centric sequences/
total words than will working class families.
The preface 'I think', an ego-centric sequence, is
probably an indication of semantic uncertainty. It invites
a further 'I think' on the part of the receiver. The
sequence signals difference and relates the sequence to the
person. It symbolizes discretion. An ego-centric sequence
also allows the listener far more degrees of freedom and may
invite the listener to develoo the communication on his own
terms. It facilitates logical develonment and exploration
of a particular area. Thus middle class families should Show
a greater tendency to employ ego-centric sequences in their
elaborated code than working class families.
H7: Middle class families will exhibit a higher
proportion of subordinate clauses/total
finite verbs than will working class families.
Pecent studies quoted in Chapter III of Lawton (1968)
agreed that 'maturity' of expression is marked not only by
an increase in the frequency of use of subordinate clauses,

but also in the complexity of their structuring. Subordinate

‘pm‘f‘m‘l

50

clauses are clearly identified with using an elaborated code
which is associated with middle class writings.
H8: Middle class families will exhibit a rreater
proportion of passive verbs/total finite
verbs than will workinm class families.
A greater pronortion of passive verbs is an obvious

example of an impersonal form which is more suitable for ”H

_‘___ —_!

less concrete writinc. Jespersen (192M) suqqests that

F ---_

passive voice may be used when there are special reasons like
tact or delicacy for not mentioninq the active subject and
where the passive may facilitate connection between one
sentence and the next. Uhese examples of the use of passive
voice are fairly sonhisticated, requiring a high dewree of
control over the language forms.

E9: Fiddle class families will exhibit a

hicher proportion of the preposition
'of'/of and in and into than will
working class families.

An elaborated code, sumoests Bernstein, is associated
with greater selection of propositions symbolizina logical
relationships than with those indicatinq spatial or temporal
relationships. mhe preposition 'of' has the lomical relation-
ship connotation as well as an adjectival quality, and thus
greater restraint is placed on this form of qualification

and use.

51

H10: Lincuistic codes of middle class families
will exhibit a hisher abstraction index
score than will the linauistic codes of
working class families.

Bernstein's thesis maintains that there is a relation
between Speech systems and orientation towards abstract
formulation, and thus 'abstract' writinq would be revealed
in middle class lanquaqe more than in workinq class communi-
cation codes. In addition, abstraction and veneralization

are usually closely connected and these two (scales) concepts

are wenerallv associated with middle class lanquaae.

Summary of Specific Hypotheses

l - Workiny class family members will exhibit a
mreater proportion of:

a-pronouns to total words

b-pronouns 'you' and 'they' to total pronouns
c—socio-centric sequences to total words
d-Cloze Procedure score completions

2 - Middle class family members will exhibit a mreater
prOportion of:

a-pronoun 'I' to total pronouns

b-eqo-centric sequences to total words
c-subordinate clauses to total finite verbs
d-passive verbs to total finite verbs

e-use of preposition 'of' to of, in, and into
f-abstraction index level

5?

Summary

Sociolinquistic research reviewed in Chapter II has
provided significant information for the development of a
set of ten hypotheses desimned to test the sociolinquistic
theories discussed in Chapter I. The linmuistic variables
used in the ten hypotheses were selected because of their
critical role in distinquishing elaborated and restricted
codes in previous research. Two of the variables, Cloze
Procedure and the Abstraction Index, were employed because
of their apparent conceptual value for Bernstein's develOp-
ment of a sociolinquistic theory of linguistic codes,
education and social class.

Chapter III will provide a description of the research
desicn, sampling procedures, operationalization of linquistic

variables, data collection and statistical analyses.

I
gr I
-

CHAPTER III

RESEARCH DESIGN

General Studv Desicn

 

The present study employed a select Vroup of 16
families in the Holt, Wichiaan, area. As no exoerimental

variables were manipulated, descriptive data were employed

1: .4 “a A
- 1.88.: 5‘ fl
4 I i

to test the hypotheses, based upon content analyses of the

linguistic reSponses obtained in the 16 personal interviews.

Sampling Procedure

 

The reSpondents selected for the present study were
16 families (from a larger sample of 25 families) with the
following characteristics:
1 - each family has a boy between lb and 16 years-
of-age;
2 - each family consists of the boy's natural parents;
and
3 - that the boy and his parents have been living
together, as a family.
The sample was limited to boys and their parents as it would
require a much larger sample to make adequate comparisons
between boys and mirls on the major linduistic variables

under study.

53

SA

Listings of families based on the above criteria
were obtained from school census records. A final list was
compiled of approximately 100 families which met the quali—
fications established by the investimator. A personal letter
was sent to each family, explaining the purpose of the study
and asking for their cooneration. (See Appendix A for letter).
After two or three days, the letter was followed by a phone
call, to set up an appointment-time for the interview.

Of the 100 families, one out of every four indicated
that they would take part in the study. Pour call backs
were made on the remaininfi 75 families, with approximately
20% of the families not being contacted. Of the remaining
60 families who were reached, most refusals were due to a
conflict with summer vacation schedules and a general
inability to get three family members toqether for the inter-
view. Interviews were conducted over a seven-week period
in the months of July and Auqust, 1968.

Of the 25 families interviewed 16 were selected to
maximize differences on criteria considered significant in
terms of Bernstein's sociolinguistic theory of elaborated
and restricted code-users. The criteria are listed in
order of importance.

1 - mother's education level

2 - father's education level

3 - annual family income

55

A - general reactions of the investiqator as to
family's job description, appearances of family's
home, etc.
These four criteria allowed a distinction between the two
samples, Vorkinm Class and Viddle Class, as follows:
Education level was arbitrarily set as 12 years of school
or less for workina class mothers and over 12 years of mud

education for middle class mothers. With this criterion

E -__Vln'J.'
I *.
1

five families (coded 01, 03, 07, 09, and 12) were placed in
the working class and four families (coded ll, 13, 15, and
17) placed in the Middle Class. Seven other mothers all

had an even 12 years of school. Employinq the criterion of
12 years of school to the father's educational level

enabled the investigator to place three more families whose
fathers had less than 12 years of school, in the Workina
Class sample. Five other families had fathers with a hiqh
school education or above and were thus placed in the Middle
Class sample. Income levels substantiated the placement

of the seven mothers who all had a hiqh school education.
The three placed in the Working Class sample had an annual
median income of $8,000, while the four placed in the Middle
Class had an annual median income of $17,000. The Working
Class and Middle Class each contained eiyht families upon
which the analyses for the present study were made. A

description of the two samples are in Table 1 below.

56

Table 1.--Descriptive Characteristics of Working and Middle
Class Samples

 

 

 

 

 

 

Working Class Middle Class
Education Level# Education Level#
Income* Father Mother Income* Father Mother
15.0 6 10 30.0 12 12
11.0 7 10 8.5 12 1”
Retired 12 11 13.0 12 16
8.4 8 12 1U.0 15 12 rmJ
11.0 12 10 15.0 15 13 5
13.0 12 10 13.5 21 16 “
7.0 11 12 20.0 18 12
8.0 11 12 12.0 16 12
MD.=9.7 11 10.5 13.7 15 12.5

 

*Income in thousands. flfducationILevel is last year
attended school. Ages of parents was exactly one year
younger for both father and mother in the Middle Class
families than in the Working Class families.

 

Definitions of Dependent Variables

 

The variables defined below are those used as
dependent variables in the ten hypotheses discussed in
Chapter 11. Definitions for all but Cloze Procedure and
the Abstraction Index are taken from definitions used by

Bernstein, Lawton or from the Revised Plain English Handbook,

 

by Walsh and valsh, 1959. The definition of Cloze Procedure
is taken from the author of the measurement technique,
Wilson Taylor (1953). Paul Gillie (1957) constructed the

Abstraction Index, a simplified formula for measuring decrees

57

of abstraction in writina. The variables or linquistic

units are defined immediately following each hypothesis.

Hypothesis I: Pronouns/total words are essentially
a type of noun that cannot take the word 'the'
immediately in front of it. Doubtful words
are tested by tryinq to insert a 'the' in a
sentence similar to the one in question.

Px.: is very sad. went away.
ExampIes of pronouns are: I, you, he, she,
it, we, they, me, yours, him, her, its, us,
them, mine, yourself, his, hers, itself, ours,
theirs, myself, themselves, this, that, such,
some, several, all, any, most, each, either,
both, few, many, none, anyone, somebody,
somethinc, everyone, nobody, no one, nothinq,
one, two, three, etc.

hypothesis IT: The pronouns 'you' and 'they'/
total pronoun were defined simply by countinq
each instance of use of the pronouns by the
subjects.

Hypothesis III: Examples of socio-centric
sequences (S.C.)/tota1 words are:
'isn't it', 'wouldn't it', 'wouldn't you',
'hadn't it', 'you know', 'ain't it', etc.
Socio-centric sequences are usually used at
the end of a sentence. Bernstein susqests
that "socio-centric sequences which are
venerated basically by uncertainty, may be
transmitted as a response of the speaker to
the condensation of his own meanincs. The
Speaker requires assurance that the messaqe
has been received and the listener requires an
opportunity to indicate the contrary.... S.C.
sequences test the ranpe of identifications
which the Speakers have in common." (1967, p. 205)

Hypothesis IV: Cloze Procedure is designed to
index the correspondence of a source's system
of lanquaye habits - includinm both semantic
and mrammatical habits — to those of another
user of the same lanauaae. A Cloze score is
a measure of dearee of ease or difficulty of a
particular messame for a particular receiver.
The score is determined by the ability of an
individual to replace items deleted from a
message.

58

Taylor has found that deletinq every 5th word,
with a random selection of the 1st word deleted
(the procedure followed by this investivator)
is a sensitive measure of readability or
comprehension. Scorinq the percentaqe of
'exact' words correctly replaced provides
the same order of difficulty level as allowina
synonyms. Taylor also susqests that scoring
'exact' words is probably more sensitive to
fine differences in style of different sources.
Only exact words replaced were scored as
'correct' in the present research.

Cloze scores were employed in this study
to draw inferences about, or make comparisons
between, elaborated and restricted code-users
with a constant set of receivers.

Hypothesis V: The pronoun 'I'/tota1 pronouns
is defined simply by counting each instance
of use of the pronoun in a subject's Speech.

Hypothesis VI: Examples of ego-centric sequences
(E.C.)/tota1 words are: 'I think', 'I believe',
'I know', 'I mean', etc. As with socio-centric
sequences, E.C. sequences are generally found
at the end of a sentence, althoumh sometimes
at the beainnina when the verb of the sequence
is not the main verb of a clause. F.C.
sequences do not usually require affirmation but
"often invites a further 'I think' on the part
of the listener," according to Bernstein.

"The sequence signals difference and relates
the sequence to the person." (1967, p. 206)
E.C. sequences allow greater freedom for the
listener and may suggest to the receiver that
he develop the communication on his own terms
with further elaboration. Both ego-centric
sequences, used more frequently by elaborated
code-users, and socio-centric sequences, used
more frequently by restricted code—users,
"play an important role in maintaining the
equilibrium which characterizes the different
codes." (1967, p. 207)

Hypothesis VII: Subordinate clauses/finite verbs
or dependent clauses are introduced by
subordinate conjunctions such as: as, as if,
because, before, if, since, that, till, unless,
when, where, and whether. According to Walsh,
a subordinate clause does not make sense when

59

standing alone, that is, it is dependent upon
other words to give it 'complete' meaning. The
subordinate conjunctions connect two clauses of
unequal rank, a dependent clause and an in-
dependent clause on which it depends. anmples
of subordinate clauses are:

l. I was here before you came.

2. Robert delivers papers before he comes to

school.
3. When I leave, I will take the lbrth road.

 

 

 

Hypothesis VIII: Passive verbs/finite verbs are
verbs of the passive voice which denote that
the subject of a sentence receives the action.
The passive always has a verb phrase composed
of a form of the auxiliary verb 'be', followed
by a past participle. Examples of passive verbs
are:

1. The dom was called by the man.

2. The letter has been written by the manamer.

 

 

Hypothesis IX: The preposition 'of', in proportion
to the use of prepositions 'of' plus 'in' plus
'into' were counted by notinm each instance in
which a subject used one of the three prepositions
in his Speech.

Hypothesis X: The Abstraction Index is described by
Gillie (1957) as a simplified formula which was
derived in part from the Flesch abstraction
formula. The correlation between Flesch and
Cillie's Abstraction Index (AI) yields a multiple
P of .8229. Gillie's AI composed of three
units is thus equivalent to Flesch's l6-unit
measure, but yet sicnificantly easier to apply
to messages.

Directions for use of the A1 are as follows:
1 - Do not use on messames less than 200
words in lenrth.

2 - Count the number of finite verbs per
200 words. Do not count any form of the
verb 'to be' when used only as a copula
to link the subject and predicate
complement.

3 - Count the number of definite articles and
their nouns per 200 words. Do not count
any definite article - noun sequences
if there in an intervening adjective or
when 'the' modifies an adjective or noun-
adjective.

tr-*-1

60

A — Count the number of nouns of abstraction per

200

words. Count all nouns endinq in the

suffixes "-ness, -ment, -ship, -dom, -nce,
-ion, and —y," includinm the plurals of such
nouns.

5 - Add
add

the numbers found in Steps 2 and 3 and
'36' to this sum.

6 - Multiply the number found in Step A by '2'.

7 - From the total found in Step 5, subtract the
result of Step 6. The result of this sub-
traction is the abstraction score.

Abstraction scores found by the above procedure
are interpreted as follows. Gillie succests that

when de

scribing the abstraction level of a messace,

one should use the verbal descriptions rather than
the numerical descriptions so as not to suqmest

the pre

0
19
31
‘43
55
67
79

ciseness often associated with numbers.

- 18, very abstract

- 30, abstract

- N2, fairly abstract
- 5“, standard

- 66, fairly concrete
- 78, concrete

- 90, very concrete

The definitions of variables given above composed the

criteria employed for content analyses of the taped dis-

cussions among three family members, of which only the

mother's and teenager's utterances were analyzed in this

study.

The fol

lowinm section briefly describes the procedures

employed for collectinq the research data.

Data Collection

 

Personal interviews were carried out in the home of

each family.

After some preliminary remarks to place everyone

 

61

at ease, the interviewer introduced the family to the
immediate task at hand, namely, their discussion of various
tOpics dealing with adolescent problems.

The discussion tOpics selected were desiqned to meet
several criteria. They had to be sufficiently interesting
to the subjects to insure motivation for communication.
They had to involve questions of social reality for which

there were no 'correct' answers. They had to permit several

"'- 1......31
i

f’

alternative solutions and to allow for differences of opinion
among family members. They had to be topics on which all
subjects would have adequate information to carry on a dis-
cussion. They had to allow for various levels of a
specificity-generality dimension or a concrete-abstract
dimension.

The interviewer read the following instructions:

"Below are a series of situations which are
discussed in most families at one time or another.
Families seem to handle the problems in different
ways. We would like for you to discuss among
yourselves as many points of View that you are
familiar with. From these different views,

select a view which represents the thinking of
your family.

Please spend some time with each of the situations.
There is no right or wrong answer...only what your
family feels is the most appropriate answer for the
situation. You will Have 30 minutes to discuss

all four situations below. Don't worry about

the time; we will inform you when the half hour

is up."

The list of sugaested discussion questions qiven to the

subjects is reproduced in Appendix B.

62

After readinq the instructions, the interviewer
answered any questions that the subjects asked. At this
point, the interviewer turned on the tape recorder and
indicated that he would leave the room until they finished
their discussion, or until the 30-minute time limit for
discussion had elapsed. Actual discussion time for three
members of the family unit, father, mother and teenacer,
ranmed from 10 minutes 30 seconds to 31 minutes u seconds,
with an averaee (mean) time of 18 minutes Q5 seconds. The
averace time for the aroups desionated Workinm Class and
Middle Class were 19 minutes 39 seconds and 17 minutes
51 seconds, respectively. The averawe time for each wroup
may not correlate with total number of words as the word
counts used in this study are only for mothers and teenaaers
and not for fathers. Upon completion of their discussion,
the interviewer handed out a brief questionnaire to each
member of the family. The complete questionnaires civen
the respondents are presented in Appendix C. Only a few
questions relevant to demorraphic information were used.
The other parts of the questionnaire were administered for
purposes of another study (Costello, 1969). After respondents
had filled out the questionnaire, the interviewer amain
answered any questions they had. This completed the data

collection stages for each family participatinm in the study.

agar-—--—-m

Data finalyses

 

Content analyses were made of ten different linguistic
units or dependent variables. In light of the non-random
sampling procedures, small samples, and the characteristics
of the variables used in the hypotheses only ordinal levels
of measurement were assumed. Thus nonparametric statistical
analyses were made. Tnour major statistical procedures were
employed:

1 - Kendall Coefficient of Concordance: N (Fiewel,

1956) was used for overall analysis of eicht
variables (variables in Hypotheses 1-2 and
5-9) by four samples (Working Class teenaders
and mothers and Middle Class teenagers and
mothers). That is, the Fendall b’was used to
test statistical differences amonc the four
samples or groups.

2 — Fruskal-Wallis one-way analysis of variance

(Siecel, 1956) was used to analyze each of
nine linguistic units (the eight mentioned
above plus the Abstraction Index) by the four
groups, that is, the fruskal—ballis was used
to test any differences amonq the four vroups

on any one given variable.

6A

3 - The Mann-Phitney U test (Siegel, 1956) was
used to analyze individual differences among all
combinations of two groups from the four samples
on each of the nine variables mentioned above.

u — A t-test (WcNemar, 1962), the only parametric
statistic employed was used to test for
statistical differences between Middle Class
families (father, mother and teenager) and
Pbrkinq Class families on the Cloze Procedure
data. Cloze Procedure scores were conducted
on one of the three discussion tOpics, 'fads'.

A section of discussion approximately A32 words
in length was selected from six of the interview
discussions, three each from family discussions
labeled bbrkinq Class and Middle Class.
Approximately 87 blanks or deleted words were

in each secment of discussion analyzed by the
Cloze method. Two subjects were randomly
selected from a University of Montana basic
speech class, to fill in the blanks on each
message. Thus a total of 12 subjects were used
on which the Cloze score is based. The percentace
of exact words correctly replaced was the basis
for the Cloze score. (See Appendix D for
specific instructions and a copy of the messages

used for the Cloze Procedure.)

"f““w
i

It:
i

65

All linpuistic counts were made by a coder trained
by the experimenter. The first three family discussions,
Coded Ol, 03 and 04, were completed by both the coder and
the experimenter. 'The coder was allowed to continue once
the experimenter and the coder obtained identical counts
on all linguistic variables. Disavreements were obtained on
'finite verbs' and 'passive verbs'. Vith further discussion .1

and clarification of criteria for these variables, agreement

.F““*"'U

was obtained. A random sampling of further discussions
and linguistic variables supmested that the coder and the

eXperimenter were in perfect avreement.

CFAPTEP IV

R FSULTR

The results of the present research are based upon
data from four samples, eiwht members per sample (a total
of 32 subjects). Linguistic content analyses were per-
formed on taped discussions of mothers and teenapers of
'Working Class' families and of '“iddle Class' families.
The average number of words upon which the analyses were
based for the four proups were: (1) workinq class mothers
(1119); (2) workinq class teenagers (515); (3) middle class
mothers (1058); and (A) middle class teenagers (531).
Bernstein's sociolinpuistic theory as well as previous
research suppested that differences in frequency of usave
should be observed for all linpuistic variables studied.
In addition, it was predicted that teenamers would exhibit
a linguistic code more similar to their mother's code than
to their peers from another social class level.
Social class distinctions were not desimned to
follow specific criteria venerally accepted in the United
States, but rather upon criteria (education levels of parents)
employed by Bernstein in his research in Eritain. Thus,
all analyses were tested for significance attx = .10 level
of confidence. However, exact probability levels obtained are
reported in the chapter.

66

r LG—m--an. 1
t

67

First, overall differences among the four groups
are given. Second, results for each hypothesis are pre-
sented which predict differences between workinq and middle
class families. Finally differences among mothers and teen-

agers of both social classes are presented.

Kendall V: Overall Analyses

The first analysis was conducted to determine if over-
all differences amonm the four Groups and eiqht linmuistic
variables (excludinq Cloze Procedure and the Abstraction
Index) are revealed. A Kendall Coefficient of Concordance:

V (Siegel, 1956) was employed for this purpose. As exnected
no significant differences (d.= .10) were found in terms of
frequency of use among the four proups. The reason was that
on three of the variables the two working class croups were
predicted to have a preater percentawe of use than middle
class groups; similarly, five of the variables predicted
greater middle class usaae than workinc class proups.

In terms of an orderinq effect, i.e., which variables
were used most frequently and which ones least frequently,
all four proups exhibited a nearly identical rankinc. A
are: 26.7“ associated with the Kendall ylwas sipnificant
beyond 1': .001 level of confidence. Thus the proportion
of prepositions 'of'/of + in + into were used most frequently
by all mroups while socio-centric sequences were used least

frequently by all groups.

68

When comparine the four proups with their usape of
the five variables on which the prediction was that *iddle
Class sroups would have a greater usape than Vorkinm Class
GPOUDS, the variance 3, associated with a Kendall V, was

sipnificant beyond“ = .05 level of confidence. Thus the

null hypothesis that the four c“roups would use the five

variables equally, could he rejected. Cnce apain the r11
orderinq of use of the variables was similar for all proups E

A

I

i
but fiddle Class teenacers used the variables most in their L.

code; Fiddle Class mothers were second, Vorkinm Class
mothers third, and Vorkinm Class teenacers used the five
variables least frequently of all the proups.

The variance 8, associated with a lendall W employed
to analyze differences amonv the four croups across the three
variables of total pronouns, pronouns you and they, and socio-
centric sequences, was not simnificantly different from zero.
Thus overall, these three variables for which Workinp Class
aroups were predicted as more frequent users than Middle
Class proups, were found not to sipnificantly discriminate
amonq the four croups.

A summary of the Kendall V analyses can be found in

Table 2.

A. 8
I...“
‘ i ‘ it

 

6Q

,

Table 2.—-kendall Coefficient of Concordance: W Analyses
of Fight Linquistic Variables Across Wour Croups, WC Mothers
and Teenagers and ”C Mothers and Teenagers.

 

Variables Sipnificances Amonm Four Groups

 

Total Pronouns
{AFC 2 MC
Pronouns 'You and They' n.s.(o(>.10)

Socio-centric Sequences 1
Pronoun 1 WC = NC

n.s. (d7.10)

Ego-centric Sequences

.3...ij

Subordinate Clauses WC) WC
8 . (0‘ =9- . 05)

Passive Verbs

Preposition 'of'

 

TV'.ann-l.--“hitneyU: Combined Pbrkinp Class vs. Combined Middle Class
Scores for Vbrkinm Class mothers and teenasers were

combined and tested for differences against the combined

scores for Middle Class mothers and teenagers. The Mann-Whitney

U test revealed the following information.

70

Table 3.--Combined Scores of Workinm Class T"‘others and
Teenapers vs. Combined Scores for Tfiddle Class Mothers and
Teenagers as Tested by the Vann-Whitney U Statistic

 

 

 

U Alpha Level
Variables statistic Attained
Pronouns/total words 103 .20
Pronouns you and they/total pronouns 85 .10*
Socio-centric sequences/total words 78 .05*
Pronoun I/total pronouns 93 .15
Ego-centric sequences/total words 87 .10*
Subordinate Clauses/total finite verbs 83 .05*
Passive verbs/total finite verbs 102 .20
Preposition of/ of plus in plus into 95 .15

 

*lndicates sipnificant differences at & 9. .10 level of
confidence.

 

Working class proups were predicted to have a c‘reater
usare index on the first three variables in the above table
than the index of Middle class groups. Support was found
for two of these, 'pronouns you and they' and 'socio—
centric sequences', but no siqnificant difference was obtained
between Working and Middle class proups on total pronouns
used.

Applying the Mann-Whitney U to variables A throuph 8,
for which Middle class croups were predicted to have preater
usage indices, reveals that only 'eqo-centric sequences' and
'subordinate clauses' indicated that combined MC proups used
a significantly creater percentape of the two variables than

combined mother-teenamer WC proups.

71

Peasons for lack of siqnificant differences on more
of the variables when combined p'roups are tested, will be
revealed more clearly when we look at differences amonq the
four proups on each variable and then between all combinations

of two sroups on each variable.

Fruskal-Vallis one—way Analysis of Variance

The Pruskal—Vallis test (Siepel, 1956) was employed
to analyze differences amonm the four proups, VC mothers and
teenamers and TC mothers and teenaqers, on each lincuistic
variable separately. Settinm sicnificance levels at d: .10
level of confidence, Fruskal-“allis (P) analyses revealed

the followinm.

Table H.--Kruskal-iallis one—way Ana1"sis of Variance Amqnc
Pour Groups (”C and ”C ”others and Teeqacers) on Each of
Tipht Variables

 

 

 

H Alpha Level
Variables statistic Attained
Dronouns/total words 6.183 .20
Pronouns you and they/total pronouns H.270 .30
Socio-centric sequences/total words 3.7A6 .30
Pronoun T/total pronouns 6.926 .10*
Eco—centric sequences/total words 2.590 .50
Subordinate Clauses/total finite verbs 6.863 .10*
Passive verbs/total finite verbs “.339 .30
Preposition of/ of plus in plus into 1.201 .99

 

*Indicates siqnificant differences at?é .IOjevel of
confidence.

 

72

Only two variables, the 'pronoun I' and 'subordinate clauses',
reached the required level for obtaining significant differences when the
four groups are tested on each variable separately. Once again some of the
significant differences between class groups is not revealed because of
either similarities between mothers and teenagers of a given class, or,
in many instances, similarities between the WC mothers' and MC mothers'
linguistic codes, as measured on the eight linguistic variables. More
clear evidence concerning the social classes and the mother-teenager
groups is revealed by making individual group comparisons with the Mann—

Whitney U test.

Mann-Whitney U test: Individual Comparisons

Forty—eight Mann-Whitney U tests were calculated which
is equal to comparisons between four groups taken two at a time for eight
linguistic variables. Each variable will be discussed in terms of sig—
nificant differences in frequency of use among the four groups. The
hypothesis associated with each variable is also given. Although the
hypotheses were stated in terms of Working Class (WC) families and
Middle Class (MC) families, results are reported for each of the two
_ groups within each class as the overall general hypothesis is concerned
with differences between classes as well as differences between in—

dividual grouPs within and between social classes.

73

prothesis l: POrkinc class families will exhibit
a greater prOportion of nronouns/
total words than will middle class
families.

The results from the Nann-Phitnev U test revealed
that ”C teenagers (MT) used a significantly greater pro-
portion of pronouns (see Table 5 below) than the second
place la’orking Class teenagers (VT) (d = .065), the third

ranked Working Class mothers (1W) (& = .05), or the fourth

ranked Middle Class mothers (”1"”) (fl = .05). I‘rLO other

1%” _ .1 FL n“ xvii;

significant differences were revealed. The results did not
support the hypothesis.
hypothesis 2: lbrkinc class families will exhibit
a greater prooortion of pronouns
you and they to total pronouns
than will middle class families.
Fesults of the tests indicated support for the
directional hypothesis. (See Table 5 below.) The ordering
of the groups in terms of most to least frequent usage, was
MT, NM, NM and MT. The only significant differences were
between VT and ET (OK = .032), and EM and MT (6‘ = .097).
Hypothesis 3: Working class families will exhibit
a greater prOportion of socio-
centric sequences/total words than
will middle class families.
Once again only partial support was obtained.
(See Table 5 below). The order of most frequent usage is

MT, WM, VT, and MM. MT was significantly different from

W? (K = .026) and ”M (CC = .018). W? was significantly

different from MM (0‘

different from M (d

Hypothesis 4 was

= .02).

74

.024), and WT was significantly

concerned with the Taylor Cloze

Procedure and is discussed later in the chapter.

Hypothesis 5:

Middle class families will
exhibit a higher proportion of
the pronoun I/total pronouns
than will working class families.

Sufficient support was obtained for this prediction.

(See Table 5 below.)

MT used the pronoun 'I' most fre-

quently followed by WT, MM and WM. Significant differences

were obtained between MT and MM (5 = .07) , MT and WM

(& = .014), and WT and WM (i = .032).

Table 5.—-Man-Whitney U Test of Significant Differences
between Pairs of Working and Middle Class Mothers and Teenagers

 

 

Variables

 

Pronouns

You & They

Socio-centric
sequences

Pronoun I

Subordinate
Clauses

Ego-centric
sequences

Passive Verbs

Preposition of

Significant Differences in Fre uenc of Usa e
MT>WT (.065); MT >WM (.05); MT>MM (.05)

(.032); WM >MT (.097)

WT4’MT

MT>WM

WT) WM

WMi’WT

None
WMi'WT

MTF'WT

(.026);

(.032);

(.052);

(.041);

(.025)

MT>MM (.02); meM (.024)
MT’MM (.018)

MT>WM (.014); MT’MM (.097)

MT>WT (.065); MM’WT (.005);
MM7MT (.052)

MT >WT (.032)

 

 

75

prothesis 6: Middle class families will exhibit
a higher pronortion of ego-centric
sequences/total words than will
working class families.

Although the order of frequency of usage was in the
predicted direction (MT, ”M, WT and WW), no significant
differences were obtained. (See Table 5 above.)

hypothesis 7: Middle class families will exhibit

a higher pronortion of subordinate
clauses/total finite verbs than

will working class families.

Partial support was obtained for this hypothesis.

PM? ‘
,

(See Table 5 above.) The order of most frequent usage of
subordinate clauses was VI, WW, MT and WT. MM was sig-
nificantly different from WT (fl = .052) and VT (K = .005).
In addition VM was significantly different from WT

(at
(d.

.052), and MT was significantly different from WT

.065).

prothesis 8: Middle class families will exhibit
a greater proportion of passive
verbs/total finite verbs than
will working class families.

General support was obtained for this hypothesis

(See Table 5 above) with the following order of most fre-

duent usage: ”T, WM, WM and WT. "T was significantly

different from WT (a .032) and ‘7‘". was significantly

different from WT (ct .0u1).

Hypothesis 9: Middle class families will exhibit
a higher proportion of the preposi-
tion of/of + in + into than will
working class families.

76

The ordering of frequency of usage was MT, TV,
tT and NM. Only the difference between MT and VT was
significant (at = .025). (See Table 5 above.)
Thus support for the hypothesis was obtained for teenagers
but not for mothers.
prothesis lO concerns the Abstraction Index and is ,
discussed below 1
‘ f“
The following chart summarizes the order of rankings
according to usage of the linguistic variables by the four “a
groups. A rank of '1' indicates that that group used the
variable most frequently, while a rank of '4' indicates

least usage of the variable. lot all rankings indicate

significant differences. (See Table 5 above.)

 

 

 

Variables Class Croups

WP if M W? T W T"
Pronouns/total words 2 3 1 N
You & They/total pronouns l 2 H 3
Socio-centric Sequences/total words 3 2 l U
I/total pronouns 2 U l 3
Ego-centric Sequences/total words 4 2 3 l
Subordinate Clauses/total
finite verbs 3 U l 2
Passive Verbs/total finite verbs U 2 l 3
Of/of + in + into 3 2 l U

 

Chart l.--SUmmary of ranks assigned horizontally to each of
four groups in terms of variable usage.

 

77

Two other variables were employed to test differences
between the four groups -- Taylor Cloze Procedure and the
Abstraction Index. The two hypotheses concerned with these
two indices are discussed below.

hypothesis H: Linguistic codes of working class

families will exhibit a higher Cloze
score than will the linguistic
code of middle class families.

The parametric t-test typically used to test Cloze .
scores was used to test differences between the combined
working class family discussions (father, mother and teenager),
and the discussions of the same family members of middle class
families. The data are shown in Table 6 below, which shows
strong support (“4 .005) for the directional hypothesis.

The Cloze score suggests that the working class discussions

were more readable, redundant or more predictable in terms of

 

Table 6.--Differences between Working Class
and Middle Class Tlamily Discussions as Measured
by Taylor's Cloze Procedure with Use of the t-test.

 

 

wc WC

 

heap Percent Correct 60.3% 47.2%

tdf=lO = 3.652; «4 .005

tflae completion scores, than were the Cloze scores of middle

clxass families.

78

Typothesis IO: Linguistic codes of middle class
families will exhibit a higher
abstraction index score than will
the linguistic code of working
class families.

lo support was obtained for this hypothesis. fill
abstraction scores fell within the range of 55 - 66, labeled
by Cillie as 'fairly concrete'. peasons for lack of sig-
nificant differences among the four groups will be discussed
in the following chapter.

The following list summarizes the results obtained from
the present research as regards two general and ten specific

hypotheses.

Ceneral Hypotheses

l - Members of families of the 'Working Class'
sample will exhibit significantly different
scores on all dependent linguistic variables
than members of families of the 'Middle Class'
sample. (Partial Support)

2 - Mothers and sons of each sample will exhibit
a more homogeneous use of linguistic variables
than will mothers or sons across samples.
(Supported)

Specific Hypotheses

l - Working class families will exhibit a greater
pronortion of pronouns/total words than will
middle class families. (fot supported)

2 - Working class families will exhibit a greater
prooortion of pronouns 'you and they' to total
pronouns than will middle class families.
(Supported)

3 - Working class families will exhibit a greater
prooortion of socio-centric sequences/total
words than will middle class families. (partial
Sunport)

‘

79

U - Linguistic codes of working class families
will exhibit a higher Cloze score than will
the linguistic code of middle class families.
(Supported)

5 - Middle class families will exhibit a higher
prOportion of the pronoun 'I'/total pronouns
than will working class families. (Supported)

b — Middle class families will exhibit a higher
pronortion of ego-centric sequences/total
words than will working class families. (Not
Supported)

7 - Middle class families will exhibit a higher
proportion of subordinate clauses/total finite
verbs than will working class families.
(Partial Support)

f5 THE. an 1. 1‘? ‘—

8 — Middle class families will exhibit a greater
proportion of passive verbs/total finite verbs
than will working class families. (Supported)

9 - Middle class families will exhibit a higher
proportion of the preposition 'of'/of + in +
into than will working class families. (Partial
Support)

10 - Linguistic codes of middle class families will
exhibit a higher abstraction index score than
will the linguistic codes of working class
families. (Not Supported)

The analysis of Specific relationships is thus com-
pleted. In Chapter V a summary and interpretation of these

results will be presented along with suggestions for further

research.

CFAPTEP V

CONCLURIONS

Introduction

 

flhe conclusions drawn in Chapter V are based upon ’
linguistic content analyses of the oral code systems of

Working Class and nfiddle Class mothers and teenagers. Data

consist of analyses of oral discourse of mothers and their
teenave sons while the family (father, mother and teenaee
son) was discussing social problems of concern to the family
and specifically with issues the teenager would or already
had encountered. The family discussed the following topics:
are for obtaining a drivers license, adult rated movies,
modern styles of dress, and teenaoe smokinv. Sixteen
families were involved in the research. Analyses were
conducted on combined Working Class (WC) scores of mothers
and teenagers versus combined Middle Class (MC) scores of
the same family members. In addition, individual analyses

examined differences amonq the four groups.

general Hypotheses

 

The first general hypothesis suggested that members

of families of the WC sample will exhibit significantly

80

81

different scores on all dependent linguistic variables than

members of the NC sample. In addition, it was predicted that

WC families would use 'total pronouns', 'pronouns you and

they' and 'socio—centric sequences' more frequently than

WC families. WC families would also reveal a higher Cloze

score than the MC. On all other variables, it was predicted

that MC families would exhibit a higher frequency of usage r14

than WC family groups. Results generally supported that

E} (h o' _-).. .p
o

prediction. Significant differences occurred in the pre-
dicted directions on five of the ten variables with two

other variables yielding highly suggestive results. One of
the ten variables, the Al, was not critical in distinguishing
among any of the four groups.

Specifically, variables 'you and they' and 'socio-
centric sequences' yielded significant differences between
WC and NC as predicted. TWO other variables, 'subordinate
clauses' and 'ego-centric sequences; in which MC was
predicted to have a higher frequency than WC, yielded
significant results. Pronoun 'I' and preposition 'of'
were suggestive in that they reached“: .15 level of con-
fidence.

Cloze procedure was one of the best predictors of
difference between WC and MC as the WC linguistic code was
more predictable (a(:: .005) than the MC codes.

The Abstraction Index did not differentiate among

any of the four groups. Two probable reasons exist for the

 

82

observed lack of difference between the two classes. First
the Abstraction Index developed by Gillie and validated
further by Haskins has only been employed with written
materials and not oral as in this study. Particular
differences between written and oral discourse may account
for the lack of significance observed in the present re-
search. One of the factors seems to be that written messages
result in a greater amount of time Spent in planning pro-
cedures and thus the result is a more formalized style of
writing than the more informal style likely employed by
family members orally discussing topics of concern to a
teenager.

The second probable reason is that the Abstraction
Index was employed on the basis of Bernstein's conceptuali-
zation that a MC code allowed for more abstract thinking
with a greater number of relationships considered among
objects. ihus one possible reason for not finding signifi-
cance is that the measures employed in the Gillie Abstraction
Index are not necessarily relevant to the ability for
abstract thinking or more complex relationships being
deveIOped. Further discussions of differences between VC
and WC will not include the Abstraction Index as it was the
only variable of the ten employed which was not discriminatory
among any of the four groups.

Seven of the remaining nine variables suggest
differences between WC and MC individuals with five of the

variables yielding significant differences.

 

83

If one looks only at class differences between teen-
agers, we observe that six of the nine variables yield
significant differences. All differences except 'total
pronouns' are in the predicted direction. Two other variables,
'pronoun I/total pronouns' and 'ego-centric sequences/total
words' were suggestive but only reached an i = .20 level of '
confidence. to differences were observed on 'socio—centric j“

sequences'.

 

Mothers of both classes seem to suggest the reason
for fewer significant overall differences between classes.
Only one linguistic variable, 'socio-centric sepuences',
was critical in distinguishing between mothers. Four other
variables (pronouns you and they, pronoun I, ego-centric
sequences and subordinate clauses) yielded differences,
though not significant, in the predicted direction.

fhus in terms of overall differences between a NO
code and a MC code, some support was found, however the
support was observed predominantly between the teenagers'
codes.

ihe second general hypothesis suggested that mothers
and sons within each class will exhibit a more homogeneous
use of the linguistic variables than will mothers or sons
across classes. This hypothesis arose from the conceptual
differences between Bernstein and Labov. Bernstein hy-
pothesizes that mothers have a greater influence on the

development of their son's linguistic code than do the sons'

811

peer group. Thus, within class differences between mother
and son should be less than the differences between sons of
different classes. Labov suggests that peer groups have the
most significant influence and thus the teenagers would have
more similar linguistic codes than mothers and their sons.

Differences between teenagers and between teens and
their mothers yielded evidence that would tend to support
Bernstein's theoretical considerations rather than Labov's
hypothesis. Teenagers from the two classes yielded similar
results on three of the nine linguistic variables and
significantly different results on six of the nine variables.
(The Abstraction Index is not considered as it seems
irrelevant as a critical variable among any of the four
groups.)

Teenagers yielded significant differences on Cloze
score, total pronouns, pronouns you and they, subordinate
clauses, passive verbs, and the preposition of. Pronoun I
and ego—centric sequences were observed in the predicted
direction, but differences were at the f z .20 level of
confidence. lb differences existed between the teenagers in
socio-centric sequences.

Mothers were different between classes on Cloze
Procedure and socio-centric sequences but equivalent on all

other variables.

85

On the other hand, WC mothers and teenagers were
similar (not significantly different) on six variables and
only different on three: Pronoun I, Subordinate clauses
and passive verbs. MC mothers and teenagers were similar
on five variables but different on total pronouns, socio-

centric sequences, pronoun I and subordinate clauses.

w

Thus it is observed that the mother seems to have a A

""'““‘"'H
i

more significant influence on her teenage son than does

F. hill.
4.

the son's peer group. The mother's influence is greater in
the VC than for the MC. Support for Pernstein's thesis

was obtained for teenagers. Peither theorist provided
hypotheses on similarities or differences among the maternal
parents between the two social classes.

The following chart suggests the relationships among
the four groups of individuals as they are ranked in terms
of frequency of use of a particular variable. A rank of
'1' indicated that the group employed that variable more

frequently than any other group.

86

 

 

 

 

Variable Classes and Groups
WT WM M T MM
Pronouns 2 3 l U
Pronouns You and They l 2 H 3
Socio-centric sequences 3 2 l M
Pronoun I 2 H l 3
Subordinate Clauses M 2 3 l
Ego-centric sequences 3 u l 2
Passive verbs u 2 l 3
Preposition of 3 2 l U
Totals 22 21 13 2M

 

Chart 2.--Group ranking in terms of usage of eight linguistic
variables.

 

The NC teenagers' code employed the variables most
frequently on all variables except 'pronouns you and they'

and 'subordinate clauses'. The sum of ranks also suggests

in addition to the social class differences between teenagers,

that WC mothers have a greater influence on their sons than
do the sons' peer groups, MT. Although the rankings do not
suggest the same for WM and their influence on MT, WM and

NT are similar on five of nine variables whereas WT and VT

are similar on only three of nine variables.

FIRTH.” ‘—
ya

87

In the following section, each hypothesis is discussed
in terms of support from combined WC groups versus MC

groups and also in terms of individual group support.

Individual lypotheses

 

The first hypothesis suggested that WC families will
exhibit a greater proportion of pronouns/total words than
MC families. Combined scores did reveal significant

differences between the two classes, however, individual

.2": . T7u3‘d.‘mm ran-1 V

group analyses suggest that NT have a significantly greater

frequency of use of pronouns than other groups. No apparent

reasons can be found for the reversal of the predicted re—

lationships with teenagers, although not with mothers. ‘
The second hypothesis predicted similar results as

hypothesis one but in terms of the proportion of pronouns

you and they to total pronouns. Both combined scores and

individual comparisons provide support for the hypothesis.

Individual support came predominantly from teenager

differences. MM appear more like WC family members as

revealed in many of the hypotheses and the chart providing

rankings of groups on each variable (p. 86).
Hypothesis three as with hypotheses one and two,

predicted greater usage by WC members. WC families should

exhibit a greater prOportion of socio-centric sequences/

total words. Although combined WC versus MC scores resulted

88

in significant differences, individual group comparisons
are tentative because of the infreouent use of socio-
centric sequences in the discussions. Support is partially
provided except for WT who were significantly different from
both mothers groups but not from WT.

iwo WM and five MW used no socio—centric seouences
while four WT and three VT used no such variable. pro-
portions of usage by the four Groups were: WT = .001;
EU = .002; W?‘ = .0025; and “M = .0005. Thus with the no
usace of socio-centric sepuences by half the sample and very
infrequent usaee by the remainder of the sample, interpre-
tations of the results would be highly suspect.

hypothesis four was the best predictor of differences
between social classes. Taylor's Cloze Procedure applied
to WC members' code systems resulted in a much greater
predictability score than when applied to bf:C members' code
systems. Further clarification has been provided to
Bernstein's somewhat vanue description of 'predictability'
as he applied that concept to NC and MC code systems.
Fven thoudh a very small sample of University freshmen
and sophomores were employed to fill in the blanks of the
experimental sample's messapes, hiehly sivnificant results
were obtained. Greater application of the Cloze technioue
in language behavior research would appear to provide

fruitful insights into sociolineuistic behavior.

 

80

1'

The remaining six hypotheses all predicted that MC
families would have a larger frequency of-use-index than
that exhibited by the code systems of WC families.

hypothesis five concerned the pronortions of the
pronoun I/total pronouns. Only partial support is provided
for this hypothesis. Combined socres were sumqestive '
(d: .15) as were individual analyses. T‘rl’l“ exhibited the s.‘
largest index score followed by WT, MW and WM. With larder
samples and mreater control of the kinds of samples selected,
the pronoun I mipht well be a sienificant variable differen—
tiating restricted and elaborated codes as employed by NC
and ”C, respectively.

Hypothesis six, the pronortion of eco-centric seouences/
total words, is interpretable much like that for hypothesis 3,
or socio—centric seouences/total words. Althouph combined
MC versus WC scores were sienificantly different in the
predicted direction, no sienificant differences occurred
among individual proups. Pank orderinn of the croups in
terms of freouency of use of ego-centric sequences was
in the predicted direction, however. Proportions of freouency
of use were very low as follows: VT = .0135; V” = .011;

MT = .019; and MM = .016. Low freouency of usaee probably

suaqests the reason for lack of sicnificant differences.

90

Hypothesis seven supaested that W, families would
use a greater prOportion of subordinate clauses/total finite
verbs than WC families combined scores supported that
prediction (d4 .05). Only partial support was provided by
individual comparisons. WW used the most subordinate
clauses and N? were second althoumh not significantly '
different from the NM scores. MT were third and sip—
nificantly greater in usave than fourth ranked WT.
Pypothesis eight predicted that MC would use a Greater E“
proportion of passive verbs/total finite verbs than would WC
families. to differences were observed between combined
scores, however partial support was provided from individual
comparisons, predominantly from teenacer differences. WT
used the passive verbs most freouently and simnificantly
more so than WT. WW, second in terms of most frequent
usage, used a significantly nreater proportion than their
teenage sons. Once aeain there exists a problem in inter-
pretation because of infrequent use of passive verbs. One
WM, three WT and one M'I‘ used no passive verbs. Proportions
of usage were very low: FW‘= .016; NM = .03U5, MT = .0M8,
and MM = .0315.
Hypothesis nine summested a yreater prOportion of use
of the preposition 'of'/of + in + into by MC than WC families.
Combined analyses were only suggestive (d.= .15). Individual

support was provided by sienificant differences observed

91

between MT and WT. One interesting observation is that
mothers were very infreouent users of the preposition 'into'
while teenagers used both 'in' and 'into' very infrequently
or not at all. Thus the proportion of 'of'/of + in + into
is greatly increased. What affect this might have overall
is not clear at this time.

The final hypothesis concerned with the abstraCtion
index was discussed earlier. Two suggested reasons for the
Abstraction Index not being discriminative among the four
groups were its application to oral rather than written
messages, and the conceptual inadequacy for employing the

variable.

Summary and Interpretations

 

In summary, seven of the ten hypotheses provided
support on the directional hypotheses although hypotheses 5
and 9 only reached the 5.1:.15 level of confidence. Individual
support for the ten hypothses was mainly gained from teen-
ager differences. Fight of the ten hypotheses received
partial support from analyses of WT and MT. Hypothesis one
was not supported in that the observed relationship was in
the opposite direction of that predicted. Hypothesis ten
suggests that the Gillie Abstraction Index may not be
applicable with the kinds of samples employed or with the

oral code systems.

Wimpy-V

92

From the data analyzed in the present research, it
appears that Eernstein's theory of sociolinguistic behavior
or restricted and elaborated code systems, is supported
generally and Specifically on many of the individual lin-
guistic variables. In that mothers of the two classes were
significantly different on only one variable, socio-centric
sequences, a slight modification in Pernstein's theory may tml

be desirable. That is, that age may in the long run over-

 

ride any differences in restricted and elaborated code systems h.
which appear at earlier ages. (Pernstein had found differences
at ages 5, l2 and 16.)

A further consideration must be made in light of the
particular design employed in this research. Mothers and
their sons discussed the topics while at home and in inter-
action with one another. Thus it is possible that both WC
and NC families would employ a common restricted code.
Bernstein's research and theory suggests that an elaborated
(MC) code user can also use a restricted code and frequently
does when in closely related groups such as families, prison
groups, and close friends. It would seem that observed
differences are all the more significant in lipht of this
conceptualization. fowever, this reasoning might place some
Qualification on the stated support for Bernstein's rather
than Labov's predictions that mothers have a greater

influence on the code system of their sons than do peer

93

groups as Labov predicts. Under the present desicn, the
researcher perhaps biased the results in favor of the
Bernstein prediction.

Eoth Lawton and Pobinson had sudmested that yiven
appropriate desicn features, restricted code users could
be forced to use "somethinp that approaches an elaborated

code". Certainly more research with carefully created

I

desicns must be employed to further test the hypothesis

that WC or restricted code users can also employ an elaborated
code. In an environment as constructed in the present
research, it was found that where both classes micht well
use a restricted code, simnificant differences were obtained.
In the first chapter the writer drew an analogy be-
tween the effect that a machine lanquaqe has upon the in-
formation processinv of a computer with that of the human
linguistic code system and its probable effect upon human
information processing. Although the present study was not
designed to determine effects of restricted and elaborated
codes on further educational potential, and in essence, one's
Weltanschununa, Bernstein (1961), provides considerable
explanation of patterns of difficulty which the workinm
class student mimht have in trying to cope with education
as it is given in our schools. These patterns of difficulty

are only suggestive from past research. Fernstein states

that we can say that the probability of finding a pattern

of difficulties is greater if the pupil's orimin is lower
working-class.

Thus a loncitudinal study is necessary in order to
determine the possible detrimental effect that a restricted
code mimht have on a pupil's information processing abilities.
Only with such studies will one be able to becin to under-
stand and thus facilitate the lower working class individual
in preparina to cope with modern society and white middle-
class school systems. And in turn one would be able to
better adapt our school systems to different subsets of
peOple in our society.

The present chapter discussed particular and general
conclusions rewarding differences between restricted and
elaborated linguistic code systems as they relate to Working
Class and Middle Class families reSpectively. Some impli-
cations were suymested in terms of further research into
socio-linquistic behaviors. The final chapter will briefly

summarize the research reported in previous chapters.

 

C Fl). T-‘T E? V I

FUWMR)’

The research reported here was based upon content
analyses of the responses of eiqht Vbrkinm-Class family
discussions and eicht Middle-Class family discussions. The
16 families all resided in the Polt, Michican area and all
teenamers, ave 1U to 16, attended a common Figh School in
Polt. Ten hypotheses concerning different linmuistic
variables were tested by develooinc prOportions for fre-
ouency of usace of the ten variables. Only the code systems
of mothers and their sons were analyzed in this study.

The ten hypotheses were developed from the socio-
linguistic theory proposed by Fasil Bernstein of the University
of London. Professor Fernstein proposes that family con-
trol patterns within Workinc Class families give rise to a
linvuistic code system (restricted) which is significantly
different from the code system (elaborated) developed in
Middle Class homes. The differences, both structural and
lexical, between the two codes mive rise to quite different
means of perceiving identical phenomena (social, intellectual
and emotional objects). A key difference in the two codes
is the decree of eXplicitness or intent which each exoresses.
An elaborated code expresses a creater decree of intent

and thus reflects a different form of the structurino of

95

 

06

',

feeling and so the very means of interaction and response to
the environment, i.e., a different means of processinm infor-
mation.

From Pernstein's sociolinouistic theory it is succested
that the linmuistic code of a UC teenacer will be more
similar to his mother than it will to his peer aroups. The
same prediction is made for NC teenaoers. Villiam Labov
of Columbia University summests the opposite, that is, that
the peer group has the createst influence on a teenager's
code system and thus the teenamers of Pbrkinq and Viddle class
societies will be more similar than the mother-teenamer
codes within the same social class.

Family discussions of social issues were taped and
transcribed. It was predicted that mothers and teenamers of
the same social class would be more similar than teenamers
from different classes. Fupport was provided for this weneral
hypothesis. It was also predicted that PC families would
differ from NC families on the variables listed in the ten
hypotheses. Apain, support was obtained.

Specific hypotheses in which WC members would exhibit
a greater prOportion of frequency of use of the linmuistic
variable than MC members, were developed for prooortions
of pronouns/total words, pronouns you and they/total
pronouns, socio-centric sequences/total words and the Taylor

Cloze Procedure. Support for the hypotheses was obtained

 

97

for all but pronouns/total words. Cloze scores yielded
the most significant criterial attribute of all variables
in distinauishinc between the two social classes.

For six hypotheses, codes of WC families were
predicted to have a preater proportion of usame than wC
members' code systems. These variables were: pronoun I/
total pronouns; eco-centric seouences/total words; subordinate

clauses/total finite verbs; passive verbs/total finite verbs;

 

preposition of/of + in + into; and the Gillie Abstraction
Index. The Abstraction Index was the only variable which
did not distinpuish between any pair of the four groups.
All other variables yielded data stronwly favorinm the
predicted direction althouph the Greatest differences were
found between teenavers and not mothers.

The data obtained for the present research provides
evidence supportinm Eernstein's socio-linguistic theory
both for differences in codes systems between social classes
and for a more significant influence of the mother, rather
than the peer group, on the teenager's code system. The
support for Eernstein's theory may be even more potent
in that his sociolinquistic theory and research was based
upon British social class systems whereas the present
research was conducted in the United States where a quite
different class system exists.

reeded research is on the refinement of linpuistic

variables, specifically structural units based upon a
generative grammar.

In addition, lonmitudinal research is urmently needed
to determine the extent to which two code systems, restricted
and elaborated, may influence educational potential and
communicative effectiveness. That is, how does the
differential use of communication codes affect human in-
formation processinq, human problem solvinm, decision—

makinm, and thinking?

~

PEWFPFNCES

Pereiter, Carl. "The Pelative Importance of Verbal and
honverbal T”actors in Cultural Deprivation: Pvidence
from Children with Sensory handicaps," (Urbana, Ill.:
Institute for Pesearch on Exceptional Children
[mimeo], 1965).

, and anlemann, Siemfried. Teachinm Disadvantaced
Children in The Preschool, anlewood Cliffs, l.J.: -1J
Prentice-Hall, 1966.

V

 

 

 

Berlo, David. The Process of Communication, Few York:
Bolt, Pinehart & Winston, Inc., 1958.

 

Bernstein, Basil. "A Public Languame: Some Sociological
Implications of a Linguistic Form", Prit. J. Sociol.,
10, 311, 1959.

 

. "Plaborated and Pestricted Codes: An Outline",
Socioloqical Inquiry Pew York: 36 (2) 1966 p. 259.
3 3 9

 

 

. "Linguistic Codes, Hesitation Phenomena and
Intellicence", Language and Speech, 5, Part 1
(October-December 1962a), pp.*3l-H6.

 

 

"Social Class, Linguistic Codes, and Grammatical
IIIements", Lanquame and Speech, 5, Part 4 (October-
December, 1962a), pp. 221-2HUT

 

 

, "Some Sociological Determinants of Perception",
4—Prit. J. of Sociolomy, 1958, p. 161.

 

 

, and Henderson, D. "Social Class Differences in the
Belevance of Lanauape to Socialization", Sociolovy,
3, (1), January, 1969.

 

 

Bloom, Benjamin 8., Davis, Allison, and Hess, Bobert.
Compensatory Education for Culture Deprivation,
Lew York: iolt, Hinehart and Vinston,'1nc., 1965.

 

Brown, Roger. Social Psycholooy, Few York: Free Press,
1965, See Chapters 6 and 7.

 

Chomsky, Foam. Syntactic Structure, The Hague: Mouton, 1957.

 

Costello, Daniel E. "Study of the Relationships Between
Family Communication Patterns and Adolescent Autonomy,"
Unpublished dissertation, Department of Communication,
Michigan State University, East Lansing, Michigan, 1969.

99

100

Dane, P. B. "The Identification and Measurement of
Environmental Process Variables That are pelated to
Educational Achievement," Unpublished Ph.D. disserta-
tion, University of Chicamo, 1963.

Gillie, Paul J. "A Simplified Pormula for Veasurinp
Abstraction in Pritins," J. of Applied Psycholovy,
Vol. 91 (A), 1957.

 

Coldman-kisler, P. 'Continuity of Speech Utterance: its
Determinants and its Simnificance," Lanmuame and
Speech, Vol. U 1961.

!

 

Jesperson, O. The Philosophy of Grammar, London, 192u.

 

Labov, William. The Study of Nonstandard Enmlish,
published by the Iational Council of Beachers of
Enmlish, 1970.

 

 

Lawton, Denis. Social Class, Lanmuaae and Education,
Poutledpe and Fevan Paul, London, 1968.

 

'\

Miller, James. 'Tiviny Systems: Basic Concepts,"
Behavioral Science, Vol. 10, lo. 3, July, 1965.

 

Fewell, A., Shaw, J. C. & Simon, P. A. "Elements of a
Theory of Puman Problem Solvina," Psycholomical
Peview, 65 (3), 1959, 151-166.

 

Piamet, Jean. The Lancuame and Thoumht of the Child,
Few York: T.‘eridian Books, 1955.

 

Pobinson, W. P. "The Ilaborated Code in Porkino Class
Lanmuape," Lanmuame and Speech, Vol. 8, 1965.

 

Sapir, P. Culture, Lanmuaoe and Personality, Univ. of
Calif. Press,’1956, p. 70.

 

Schatzman, Leonard, and Strauss, Dnselm. "Social Class
and Modes of Communication," The American Journal of
Sociolomy, 00:329-98, 1955.

 

 

Shuy, Power W. "Ponnie and Clyde Tactics in Enclish
Teachinm," The Florida W. L. Deporter, V. 7 (1),
1969. A paper presented at the 1968 PCUEiConvention
in Milwaukee.

Taylor, W. "Cloze Procedure: A new tool for measurinm
readability," Journalism Duarterly, 30, 1953,

Oh. “IS—“33.

 

101

. In rfrends in Content Analysis (ed.) Ithiel de Solo
Pool. U. of Illinois Press, 2nd. ed., 196“, pp. 78-88.

lhayer, Lee. 'Tommunication and Orcanization Theory," in
Puman Communication ”heory: Criminal Fssays
(ed.) Prank, FIX. Dance, Tew Vork: Lolt, Pinehart
& Winston, Inc., 1967.

Watzlawick, Paul, Peavin, Janet, and Jackson, Don. Prammatics
of Iuman Communication, Pew York: k3 Eh lbrton &
Company, Inc., 1967.

 

 

Phorf, B. L. Lanmuame, Thouvht and Beality, Iew York: 1 1
ia-‘iley 8. Sons, 1956. F1
2

 

L ETT H?

APFPI‘D IX A

PUT TC) RFSFUI‘JDI'NI‘S

 

103

LETTER SENT TO RESPONDENTS

July, 1968

Dear Mr. & Nrs. :

 

Most of us recognize that what we say to others is pretty
important. We are interested in the important communication
within families; that is, what do families talk about and how

do they talk about certain topics. We would, therefore, like
to study the communication which goes on between members of a A
family. *T‘

We are graduate students in the Department of Communication at
Michigan State University. Our interests are in lookina at h+
patterns of discussion within families as they relate to other ‘
kinds of social activity, such as, where you so to look for

ideas about particular tooics, and how you so about making

decisions about things of concern to the family.

 

Your family (father, mother, and teen—ape son) would be asked

to participate by sittina and discussina a few tOpics, such

as school activities, dating, and so forth. This discussion
would take about one half hour followed by a short questionnaire
concerninq how you make the daily decisions of concern to your
family. You can be assured that the answers of any Specific
family will not be made public. We are interested in mroups

of families only, and in their patterns of discussion.

In two or three days we will be contactino you by phone to ask
for your coOperation in discussinm among yourselves, tooics
concerned with areas of interest to you and your family. We
hope that you will amree to help us in our wraduate promrams.

Thank you for your time in readinq this letter. We hope that
you will be interested in our research and find an hour in
your day in which to help us.

Sincerely,

 

 

Duane D. Pettersen Daniel F. CostelIo
flraduate Assistant Graduate Assistant

AI‘FFPL‘D IX B

 

LIST OF DISCUSS ION CUFE‘TIOI’S

(l)

(2)

(3)

(L!)

105

DISCUSSION OUESTIONS

Below are a series of situations which are discussed in most
families at one time or another. families seem to handle the
problems in different ways. we would like for you to discuss
amonm yourselves as many points of view that you are familiar
with. Tfrom these different views, select a view which repre-
sents the thinkinm of your family.

Please spend some time with each of the situations. There are
no right or wrong answers...only what your family feels is

the most appropriate answer for the situation. You will have
30 minutes to discuss all four situations below. Don't worry
about the time; we will inform you when the half hour is up.

 

It has been proposed that the minimum age for setting a drivers'
license be raised from ace 16 to ace 18. Discuss the different
points of view that you are familiar with on this tooic, and
select one which represents the thinkinm of your family.

Some parents feel that if their teen-ave son or daughter wants
to smoke, the choice should be left up to the teen—aver. Other
parents insist that their teen-are son or daughter may smoke
only when the parents consider them ready to smoke. Discuss
the different points of view that you are familiar with on

this topic, and select one which represents the thinkinm of
your family.

Some parents allow their teen-ace son or daumhter to no to movies
based on violence, sex, and other adult themes. Other parents
regard films recommended for adults only as strictly off-

limits, until the teen-acer is older and more mature in his
thinking. Discuss the different points of view that you are
familiar with on this tooic, and select one which represents

the thinkinm of your family.

 

Some parents feel that their teen-age son or daumhter should be
allowed to wear their hair or clothes in keeping with the current
fads. Other parents insist that conformity to teen-age fads

is unnecessary and that teen-avers should leave the final
decision up to the parents. Discuss the different points of
View that you are familiar with on this topic, and select one
which represents the thinking of your family.

 

APPENDIX C

 

CUEETIONNAIPEF

107

CHILD'S QUESTIONNAIRE

To begin with...what specific tOpics are discussed most often
in your family at meal-time?

 

 

 

 

I'd like you to read each of the following questions carefully...
then place an "X" in the blank in front of the answer you
consider most apprOpriate. Please choose only one answer for
your father and 223 answer for your mother for eaah question.

11-12 About how often would you say your parents ask you
for your opinion on family decisions?

 

 

 

 

 

 

MOTHER PATHPP
Several times a week Several times a week
About once a week About once a week
Once or twice a month Once or twice a month
Less than once a month Less than once a month

 

 

13—1A If your parents said that they depended a great deal
on your Judgment regarding family decisions, would
you believe them?

 

 

 

 

MOTHER FATHER
Yes Yes
I guess so I guess so
Probably not Probably not
No No

 

 

 

15-16 Compared with other teen-aaers...are you more likely,
or less likely...to be asked by your parents for
Opinions on family decisions?

MOTHER FATHER
More likely More likely
Less likely Less likely
About the same About the same

17-18 In regards to family decisions, would you like to think
your parents consider your Opinions?

 

 

 

NOTHFP FATHFP
Yes Yes
I guess so I guess so
Probably not Probably not

Do lb

19-20

Now,
teen—
the p

If yo

certain as to what choice to make, how often would you look for . .

21.
22.
23.
2M.
25.
26.
27.
28.

If yo
what

29.

108

When you discuss family decisions with your parents,
what part do you usually play?
MOTHER FATHER

I talk mostly I talk mostly

I mainly listen I mainly listen

A little of both A little of both

 

I would like for you to read about some problems that
agers your are sometimes face. Listed below are some of
ossible ways of finding answers to these problems.

u were concerned about findinc a part-time iob and un-

 

OV"VFF' HOLY & ””TF“‘ PFIJ‘O“? PWWVFD

 

 

IDPAS from newspapers

 

 

IPPA9 from your mother

 

 

IDEAS from television

 

 

IDEAS from others your ave

 

 

IDEAS from marazines

 

 

IDEAS from your father

 

 

IDPAS PROM MOVIES

 

 

IDFAP from other relatives

 

 

 

 

u needed to buy clothes for school and uncertain as to
choice to make, how often would you look for . . .

OFVVFDT NF“? & WHIP}? FETJWWV NPWTFP

 

 

 

IDEAS from newsnaners

 

 

IDEAS from your mother

 

 

IPPAS from television

 

 

IDPpF from others your ace

 

 

IDEAS from maaazines

 

 

IDEAS from your father

 

 

IDEAS from movies

 

 

IDIAS from other relatives

 

 

100

a

37. On matters such as jobs and clothinn, if you could ask
for ideas from only one of your parents, which parent
would you ask?

Mother Father

 

 

Dow here is a different kind of Question. Place an "X" in
the blank in front of the answer you consider most appronriate.

38. If you had just boucht new clothes for schoo; how sure
would you be that you had made the best choice possible?

Very sure
Sure
Not sure

Ti

 

g
1‘
:1
1?

«I?

39. If you had just joined a new club at school, how sure
would you be that you had made the best choice possible?

 

Very sure
Sure
Not sure

 

 

NO. If you had just accepted a part-time job for this
summer, how sure would you be that you had made the best
choice possible?

Very sure
Sure
Dbt sure

 

Ml. Pinally...just a few more questions about yourself...
what is your age?

 

M2. And what was the last arade you completed in school?

“3. Are you taking or going to take a college prep or
vocational or business traininm courses in himh school?

College prep.

Vocational
Pusiness
uu. what subject is easiest for you?

 

M5. What subject is hardest for you?

 

46. For the first 10 years of your childhood, what state
or country did you live in for most of these years?

Did you live primarily in a rural (farm) or urban (city)
area during this time? rural urban

“7.

110

 

Do you speak a languaqe other than anlish? Yes No
that language?
Does anyone else in your family speak it? Mother Father

Is it used frequently in the home? Yes

Pb

M4
1.2:“? h
r

lll
MOTHER'S nUESTIOIVNAIFE

To bedin with...what specific topics are discussed most often
in your family at meal-time?

 

 

 

 

 

I'd like you to read each of the following questions carefully...
then place an "X" in the blank in front of the answer you
consider most apprOpriate. Please choose only one answer for
each question. ___

In general, how are most decisions made between you and your
teename son on the following situations . . .

 

ll. If your son needed new clothes for school...does he
usually choose his own, or do you decide for him?

I always decide for him

I often decide for him

Every now and then I decide for him
I seldom decide for him

I never decide for him

 

 

12. In rewards to your son's friends...does he usually
choose his own, or do you summest who they should be?

I always suggest to him

I often squest to him

Fvery now and then I suecest to him
I seldom susmest to him

I never suggest to him

 

 

13. When your son goes out with others his own ave...does
he usually come home when he wants to, or do you usually
remind him of what time to be home?

I always remind him

I often remind him

Every now and then I remind him
I seldom remind him

I never remind him

 

 

 

1U.

15.

l6.

17.

19.

20.

21.

112

In rewards to your son's datinm...does he usually
determine how often he goes out, or do you tell him
when he is allowed to date?

I always tell him

I often tell him

Every now and then I tell him
I seldom tell him

I never tell him

Finally...a few questions about yourself...what is your
ave?

 

And what was the last wrade you completed in school or
college?

Pow many children do you have livins?
Lhat are their aces?

Does anyone else live with your family? tho?

 

Do you speak a lanvuave other than lnmlish? Ves To

that lanmuame?

 

Does anyone else in your family speak it? husband
Teen-ave Son

Is it used frequently in the home? Yes lo

 

For the first 10 years of your childhood, what state or
country did you live in for most of these years?

 

 

Did you live primarily in a rural (farm) or urban (city)
area durinq this time? rural urban

 

What is your family's religion?

Protestant
-——Catholic
_Jewi sh
———Other

 

113

FATE-EFT? ' FT DU P F'I‘IO IV 14A IF I

To bemin with...what Specific tonics are discussed most often
in your family at meal-time?

 

 

 

 

 

I'd like you to read each of the followinw questions carefully...
then place an "X" in the blank in front of the answer you
consider most appropriate. Please choose only one answer for
each question. ——_

In general, how are most decisions made between you and your
teename son on the followinr situations . . .

ll. If your son needed new clothes for school...does he
usually choose his own, or do you decide for him?

I always decide for him

I often decide for him

Every now and then I decide for him
I seldom decide for him

I never decide for him

12. In rewards to your son's friends...does he usually
choose his own, or do you suaqest who they should be?

I always suwrest to him

I often summest to him

Eyery now and then I sumwest to him
I seldom sumeest to him

I never supmest to him

13. When your son aces out with others his own awe...does he
usually come home when he wants to, or do you usually
remind him of what time to be home?

I always remind him
I often remind him
Eyery now and then I remind him
I seldom remind him
I never remind him

 

 

l“.

15.

l6.

17.

18.

19.

20.

11a

In rewards to your son's datina...does he usually
determine how often he woes out, or do you tell him
when he is allowed to date?

I always tell him
I often tell him
PVery now and then I tell him
I seldom tell him

I never tell him

 

 

Finally...a few questions about yourself...what is your
awe?

And what was the last mrade you completed in school or
colleme?

 

that kind of work do you do or usually do?

 

Phat type of business or industry do you work for?

 

What was your family's approximate total annual income
for last year (1967)?

 

Do you speak a languawe other than Inwlish? Yes Po

Vhat languaee?

 

Does anyone else in your family speak it? Wife
Teen-age son

Is it used frequently in the home? Yes No

For the first 10 years of your childhood, what state or
country did you live in for most of these years?

 

 

Did you live primarily in a rural (farm) or urban (city)
area durinm this time? rural urban

 

APFflflUDIX D

 

INSTRUCTIOFS ALD SAMPLE MESSAGE

FOR COFDUCTING TAYLOP'F CLOZE;PROCIDUFF

116

IYTPODUCTIOK FUD ILFTPUCTIOIP

Introduction

 

I'm conductine some research that to me is very
important and I hope that you will consider it important
and thus do your best on the materials that I'll hand out
in a moment. It will take approximately 15 minutes to
complete the exercise.

how, many of you probably know that the nature of re—

search is such that if I told you beforehand what I was tryinp

 

to determine or what I was studyinm, that information mimht
bias or influence your behavior and thus the results of this
study. Psycholomical research shows that this is the case.
In other words, you miwht try to help me out or hinder me

in obtaining fairly accurate behavior on your part. I will
explain in detail, afterwards, exactly what I'm tryinc to
find out and how it might be relevant to you.

OKAY?

Instructions

 

If I said 'Ped, white and .' What would you say?
how if I said 'Chickens cackle but ouack,' what would
you say?

OK. That's the task I want you to do. You all have
different messaves. Work rapidly but try to fill in the
exact word that you think is missine.

Any questions?

Begin.

fC-O7

 

117

 

 

 

 

 

 

I don't _____ they should be allowed wear their
hair as as they want to. clothes in the
current . I think when you're _____ to school
you should dressed neat, in resoect.

all your lonv hair ______lice in it.
Well you really think about it's really
stupid to 1onm hair and clothes _____ if you
start out _____ yourself halfway neat and _____ go
throumh school that and maybe you so colleqe
that way, if _____ want to met a ______with a biq
company _____ something they're aonna look you
before they look _____ anybody that's a beatnik _____
somethinv.

That's more lomical that's more practical. That's
Uncle Sam wives you hair cut when you

in the Army. I know about the final beinw

 

 

left up to parents, I think they should
to a decision tomether.

right.

 

In the lonm you'll probably end up better

friends, too because

, some of your lonm friends wouldn't amree

 

with there.
how, usually the that are dressed neat

usually more intelliwent anyways.

 

 

 

AC-O7
118

A: most of the kids here don't have real

hair, not what you'd lone for colleve kids.

kids wear their hair "' compared to them.

 

 

 

 

 

P: Wearing hair and dressine like , it takes
so much their thought away from else. They
Spend too of their thouyhts on they dress.
C: And it costs a lot more
B: Surprising if they take , look at the money

save on soap.
A: I'll the soap.
B: They spend that much on deoderant they

can stand themselves.

 

 

 

 

 

A: , they don't. They don't ______bother with the
deoderant of them. 1b, I _____ that, I don't
think can insist on conformity .

B: Well, the thins is _____ in a lowical way,
should be all right. why I'm in favor

 

droppinm the driver's age to lb.

 

 

C: Actually when going into fads and , you're
changing your mind the time. You're chanminw
what everybody else is theirs too. Whenever

you your job the employers want somebody

 

that isn't to change his mind. be wanting
you for ideas and if you changing then
you're not use to them. So qonna droo you

for better.

   

mllllflfllfifllifilHIJIIIEIIIJWIEIIIEWIWEs