‘-C - c'v-o ...--- 909-. 91.... BERNSTEINIS ELABORATED AND RESTRICTED LANGUAGE CODES AS A ”FUNCTION OF SOCIO-ECONOMIC CLASS. RACE AND SEX Thesis for the Degree of M. A. MICHIGAN STATE UNIVERSITY" LARRY JEROME DeNEAL 1972 ~ n.-+.. .I..-.l-a:.I. ; U. I a! «run. .3 . . 3%...- .(oP' ,0 ,‘E 1' .I ' O I . .'.f o ‘ A In! I. , fl! 1,. [.1 o .1 Ju. . . I .y. rHEbIS J III III IIIIIIIIIIIIIIIIIIIIIIIIIIII 3 1293 01103 9991 _ '- ‘3‘3: ash. .‘- if - n . :3 a IF 9‘ I I": E1927; 'w- ..- ‘T‘-“ 419.1 - .— " :1. I i p . ..O J . -': A‘ ’o i ' ?’ ‘31 ————— 4 fl dc} Hgldw 'r-r.. r— T‘- __A ‘ —— . , INC. LIBRARY BINDERS gameronulcmenu ‘ x :/ # MSU LIBRARIES “ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. Not; 9 8.1999: ABSTRACT BERNSTEIN'S ELABORATED AND RESTRICTED LANGUAGE CODES AS A FUNCTION OF SOCIO-ECONGMIC CLASS, RACE AND SEX BY Larry Jerome DeNeal This thesis investigated the Bernstein hypothesis of- restricted versus elaborated language codes as applied to an American pOpulation consisting of Black and White males and females from four major socio—economic groups. The forty language samples used in this study were obtained from the Detroit Negro Dialect Study by Shuy, Wolfram, and Riley (1968). Bernstein listed ten discrimi- nating measures of restricted and elaborated codes. Eight of these ten measures were employed in this research and are as follows: Pronouns/total words, Pronouns ”you" and "they" per total pronouns, Socio-centric sequences/total words, Ego-centric sequences/total words, Subordinate clauses/total finite verbs, Passive verbs/total finite verbs, Number "of's"/"of” "in" and "into.” These eight measures were investigated relative to the factors of race, sex, and socio-economic status. Larry Jerome DeNeal It was concluded from the results of this research that only four of the eight measures were capable of dis- criminating between a restricted code or an elaborated code. When considering these four discriminating measures, only two were related to race, and both of these are not independent measures, but rather are dependent upon the interaction of sex and socio-economic status. The results of this research showed that only one out of the eight hypotheses Bernstein formulated was verified; therefore, it was concluded that the Bernstein model of a restricted language code versus an elaborated language code is not applicable to an American p0pulation consisting of Black and White males and females from various socio-economic classes. Accepted by the faculty of the Department of Audiology and Speech Sciences, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the Master of Arts degree. Thesis Committee: r/J(11252§Z§f;é(%£292~c\g Director Daniel S. Beasley, P 4D.‘\\ am: @ £6216“): Erwin P. Bettingfiaus, Ph.D. A.“ Herbénfir3. gyer, Ph.D. ' if? ? ..- "Iv/[c "" ""“ / _.- \ C; c/ 2/1 (go-r, use-Iv L. Eudora Pettigrew, thD. BERNSTEIN'S ELABORATED AND RESTRICTED LANGUAGE CODES As A FUNCTION OF SOCIO-ECONOMIC CLASS, RACE AND SEX BY Larry Jerome DeNeal A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Audiology and Speech Sciences 1972 ACKNOWLEDGEMENTS This writer wishes to express-his appreciation to Dr. Daniel Beasley, who has not only been an advisor but also a friend. I would also like to thank Mrs. Daun Beasley for her patience with us both. A special thanks is extended to my dear friend, Patricia Larkins.for her inspiration and assistance in the preparation of this thesis. I also wish to thank Dr. Roger Shuy for his invalu— able assistance in preparation of the materials for this thesis. ii CHAPTER II. III. IV. V. TABLE OF CONTENTS LIST OF TABLES. . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . . INTRODUCTION. . . . . . . . . . . . . . . SOCIOLINGUISTIC THEORY OF ELABORATED AND RESTRICTED CODES . . . . . . . . . . . General Sociolinguistic Theory . . . . Summary and Statement of the Problem . EXPERIMENTAL PROCEDURES . . . . . . . . . subjects . . . . . . . . . . . . . . . Transcription Procedures . . . . . . . Linguistic Analysis. . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . Effects Associated with Race . . . . . Effects Associated with Socio—economic Status. . . . . . . . . . . . . . . Effects Associated with Sex. . . . . . Intercorrelations of the.Measures. . . DISCUSSION. . . . . . . . . .,. . . . . . Implications for Future Research . . . SUWMARY . . . . . . . . . . . . . . . . . LIST OF REFERENCES . . . . . . . . . . . . . . . APPENDIX--WEIGHTED AVERAGE SCORES FOR EACH MEASURE EMPLOYED AS A FUNCTION OF RACE, SEX, AND SES AND THEIR RESPECTIVE INTERACTIONS. iii Page iv 24 27 27 29 31M 33 34 35 36 37 45 50 54 56 59 LIST OF TABLES TABLE Page 1. Distribution of social status scores in rela- tionship to the four social class labels. . . . 29 2. Correlation matrix for eight Bernstein measures of restricted and elaborated codes. . . . . . . 44 iv LIST OF FIGURES FIGURE 1. The percentage of distribution of Black and White subjects over the four socio-economic levels. . . . . . . . . . . . . . . . . . . . Mean percentage of subordinate clauses per total finite verbs as a function of Race. . . Mean percentage of preposition “of" per "of" "in" and "into" as a function of Race as it interacts with Sex. . . . . . . . . . . . . . Mean percentage of subordinate clauses per total finite verbs as a function of SES . . . Mean percentage of ego—centric sequences per total words as a function of Sex as it inter- acts with SES . . . . . . . . . . . . . . . . Mean percentage of pronouns "you" and "they" per total pronouns as a function of SEX . . . Page 30 39 40 41 42 43 INTRODUCT ION Recent theoretical speculation and research in the area of sociolinguistics by Basil Bernstein (1962) as it re- lates to language codes of teenagers of different racial, ethnic, socio-economic, and sex differences is the subject of this thesis. Bernstein's two linguistic codes are identified with lower—working class and middle-class families' means of communicating and are based upon family interaction patterns within the social classes. The two codes are dis- tinguished on linguistic, syntactic, and lexical, as well as a psychological, levels. Linguistically, a restricted code user selects from a smaller set of syntactic and lexical Options than does an elaborated code user. According to Bernstein, psychologically an elaborated code-user expresses a greater degree of "intent" or specificity than does a restricted code user. The sociolinguistic theory of Bernstein suggests that working class teenagers will exhibit a restricted code sys— tem and that a middle class teenager will employ an elabor- ated code. .Bernstein's original research (1962) at the University of London utilized sixteen year old, Caucasian males, of different socio-economic groups. However, when Bernstein's original theory was interpreted by American sociolinguists, educators, and child psychologists, it was generalized to include members of both sexes, different age groups, peeple from different racial and ethnic background, including Black Americans. Such theorists and researchers (Deutsch, 1967; Bereiter and Englemann, 1966; Hess and Shipman, 1965; and Raph, 1965) implicitly equated Bernstein's restricted code with a type of cognitive deficit. Bernstein (1970) Objected to such a relationship. Several of the researchers who proposed the cognitive deficit theory designed language enrichment programs to teach lower class Black children to speak in what Bernstein terms an elaborated code. Labov (1970) and Johnson (1970) attacked the reviews of deficit theorists on the grounds that the speech samples the deficit theorists collected were not truly representative of the linguistic behavior of Black children. In addition, their interviews with the children were not controlled for topic. Further the ability of a non-native speaker of Black dialect to accurately transcribe Black dialect was questioned. To date there has been no adequately controlled empirical investigations in the United States which corre- late Bernstein's restricted and elaborated codes with race, sex, and differences in socio-economic groupings. Thus an investigation as to the application of Bernstein's theory to Black Americans, including both males and females, appears warranted. An investigation of the "restricted" and "elaborated" language codes as they are correlated with different racial groupings, sex differences, and socio-economic classes, was carried out. This investigation utilized the data and language samples collected in the Detroit Dialect Study (Shuy, Wolfram, and Riley, 1968). The transcriptions of the taped interviews were performed by a Black male native bi-dialectic speaker of "Black lower—class dialect" and "Standard American English." These transcriptions of black and white speakers were subjected to eight of the nine linguistic measures Bernstein reported were indicative of a restricted and elaborated language code. The results of the data analysis were correlated with the three variables of sex, race, and socio—economic class. This thesis is organized into the following chapters: Chapter I will examine current sociolinguistic theory and current thought concerning "lower class Black dialect" versus "standard American Dialect." This thesis reviews related experimental research and speculation concerned with Bernstein's theory. Specific Hypotheses for the present study are developed. Chapter II provides an explanation of the study's design, sampling procedure used by the Detroit study, explanations of the sociometric measures used, and statistical analysis. Chapter III presents the results. Chapter IV discusses general and specific conclusions and implications for future research. In conclusion, Chapter V provides a brief summary of the research reported in this study. CHAPTER I SOCIOLINGUISTICS THEORY OF ELABORATED AND RESTRICTED CODES General Sociolinguistic Theory Over the last two decades a number of economic, politi— cal, and social factors have combined to bring to the fore- front of public attention the conditions of underdevelopment among human beings in all parts of the world. In response to these concerns most fields in the behavioral sciences have developed disciplines which study techniques for assess— ing and dealing with the problems of these disadvantaged populations. Sociolinguistics has emerged as a viable field of study of basic processes of communication and its regula- tive function within and between identifiable social groups. The work of investigators such as Young (1941), Irwin (1948), Anatasi (1958), and Templin (1958) has shown that the structure of the language system which an individual acquires is related to background features associated with social class, sex, age, race, and education. These authors stress the relationship language acquisition has to family inter— action patterns and the "quality" of the language used in the home. For educational purposes, Bereiter and Englemann (1966) and Martin Deutsch (1967) have equated cultural deprivation with language deprivation. Language deprivation, according to these authors, is primarily a failure to learn the uses of language. The problem with one who has a language deficit concerns the inability of the individual to encode and decode information his environment provides, i.e., to use language as a primary information processing system. Hymes (1967, p. 8) noted that "language is code, speech is message, i.e., we refer to the act of speaking but to the structure of language. According to Hymes, language is a set of rules which will generate a number of speech messages. Recent research in the field of sociolinguistics, edu- cation, and child psychology (Deutsch, 1967; Bernstein, 1962; Petersen, 1970; Lawton, 1968; Robinson, 1965) have suggested that a direct relationship exists between the language pro— cessing abilities of a child and the social class level to which he ascribes. Bernstein's early writing during the late fifties labelled his two language codes as "Public" versus "Formal." In subsequent years (1961), he chose to re-label his language codes as "Restricted" versus ”Elaborated." However, as he pointed out (1961), the two sets of code names are synonymous. "Public" code refers to "restricted" code and "Formal" code is the predecessor of "elaborated" code. According to Bernstein, one's particular social relation- ship with his environment will determine the form the indi- vidual's linguistic codes will assume. Central to this theme is that change in the form of certain relations act selectively upon principles that control both syntactic and lexical Options. These language codes create different orders of relevance and relations for the encoder and orientations for the decoder; thus one comes to learn his role through communication. This theory is, in fact, the inverse of the Whorfian Hypothesis (1956). Whorf felt the basic and primary force creating different ways of viewing and interpreting the world is communications, whereas, Bernstein views social relationships as the primary determinate of the form a com— munications code will assume and thus define the context of man's environment. These two linguistic codes, restricted and elaborated, are initially differentiated on syntactic, semantic, and lexical Option bases. Bernstein (1959, p. 311) posits the following criteria as characteristics of one who chooses to use a restricted language code: 1. Short, grammatically simple, Often unfinished sentences, a poor syntactic structure, with a verbal form stressing the active mood. 2. Simple and repetitive use of conjunctions (so, then, and because). 3. Frequent use of short commands and questions. 4. Rigid and limited use of adjectives and adverbs. 9. 10. Infrequent use of impersonal pronouns and subjects (one, it) . Statement formulated as implicit questions which set up a sympathetic circularity, e.g., 'Just fancy?’ 'It's only natural, isn't it?’ 'I wouldn't have believed it?’ A statement of fact is often used as both a reason and a conclusion, more accurately, the reason and conclusion are confounded to produce a categoric statement, e.g., 'Do as I tell you', 'Hold on tight,' 'You're not going out,' 'Lay Off that.’ Individual selection from a group of idiomatic phrases will be frequently found. Symbolism is of a low order of generality. The individual qualification is implicit in sentence structure; therefore, it is a language of implicit meanings. It is believed that this fact determines the form of the language. Bernstein states the pure form of a restricted code is seldom found but it would manifest itself in some type of ritual or ceremony. Bernstein (1958, p. 312) described the elaborated code user as one whose language displays the following character— istics: l. 2. Accurate grammatical order and syntax regulates what is said. Logical modification and stress are mediated through a grammatically complex sentence structure, especial- ly through the use of a range of conjunctions and subordinate clauses. Frequent use of prepositions which indicate logical relationships as well as prepositions which indicate Spatial contiguity. Frequent use of the personal pronoun 'I' and the impersonal pronouns, 'it' and 'one.' 5. A discriminative selection from a range of adjective and adverbs. 6. Individual qualification is verbally mediated through the structure and relationships within and between sentences. 7. Expressive symbolism discriminates between meanings within speech sequences rather than reinforcing dominant words or phrases, or accompanying the sequence in a diffuse generalized manner. 8. A language use which points to the possibilities inherent in a complex conceptual hierarchy for the organizing of material. Bernstein postulates the discriminating characteristics of restricted and elaborated codes on the basis of the follow- ing psycholinguistic theory which was written using the restricted code as a referent. The short, grammatically simple, syntactically poor sentence which is the typical unit of a public language does not facilitate the communication of ideas and relationships which require a precise formulation. Inherent in its crude form and length the sentence structure of the restricted code is not conducive to the description of processes. An approximate verb or synonym of a lower logical order may be used in the expression, while the verbal construction may place the process in an inappropriate time as a result of an insensitivity to tense in expressed communication. For example, when questioned repeatedly, the individual will retort with "that's what I meant" which gives credence to the notion of implicit meaning. Because the restricted speaker tends to emphasize 'things' rather than logical processes, 10 one would expect to find a higher percentage of nouns than verbs, according to Bernstein. The simple sentence con- struction limits the use of conjunctions; thus, conjunctions are not used to link logical thought or stress relationships between statements and reason. Stress and logical modifica— tion in a restricted code is expressed in what Bernstein terms 'crude' extra—linguistic manners, such as pitch varia- tion. The grammatically simple sentence relies heavily upon a few conjunctions (and, so, then, because) usually used in- appropriately in place of a more logical and exact distinc- tion. The approximate term will then become the equivalent for that speaker of the appropriate term. The rigid and limited use of adjectives and adverbs according to Bernstein will result in a more generalized description of processes and Objects in the restricted speaker's environment. He may perceive the elaborated codes perception of identical phenomena as containing irrelevant variables. The effects of the restricted code speaker's not making frequent use Of the impersonal pronouns (one, it) is that in substituting 'I', 'we', 'us', the speaker's realm Of expres- sion is limited to personal or shared experience. Using 'one' or 'it' involves reaching beyond the immediate experi— ence of the individual and Objectifying so as to transcend to a theoretical level of conceptualization. Bernstein (1958, pp. 313-314) states: 11 However with a public language it is much more prob- able that the pronouns 'we' or 'you' will serve an apparently similar function to 'one.' It is in fact not similar, nor is it a simple substitution; for 'we' and 'you' refer to local experience, the local social relationships and the immediate normative arrangements, and are bounded by the personal. Syntactically and semantically the restricted code user has the tendency to confound the reason and conclusion. This results in the speaker making a categoric statement which is accepted by the listener on the basis of the speaker's relationship to the listener, instead of on the basis of a logical rationale. Bernstein (1958, pp. 163—164) provides the following example of the difference between working class and middle classzmother—child interactions. The pairs are on a bus and the children are sitting on their 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: Why? . Mother: If you don't you will be throWn forward and you'll fall. Child: Why? 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. 12 Although the example is short and somewhat concrete, it is a vivid example of some of the features of the two language codes. When comparing them on a syntactic and lexical option basis, the elaborated code using mother tended to use longer and more complete sentence structures. Semantically, one mother in controlling her child placed more emphasis upon language because she wishes to make explicit and to elaborate for the child certain rules and their consequences. Socio- logically, it can be argued that the restricted code using mother and child tended to adhere to a more status based relationship, wherein the giving of a reason for a command is not a relevant variable. For restricted code users, reason is implicit by virtue of the communicating participants' social positions. The child accepts the request because one of a higher status made the request. The restricted code using mother places less emphasis upon language in controlling her child; she deals with only the particular act and does not relate to general principles and their reasoned basis and consequences. According to Brown gt al;§ (1965) theory of language develOpment, expansion of language by the mother is very important. Providing sequencing and sets of reasons for the child's 'Why?‘ 'How come?‘ etc., questions, are vital to the child's growth, as is immitation, with slight modifications which provides 'correct' grammar plus some of the function words, and auxillary words, etc. Thus it can be seen that the restricted code using mother, in effect, l3 terminated conversation at an earlier point in conversation than the elaborated code using mother. Bernstein notes that the restricted code is character— ized by sympathetic circularity. He defines this as the repetition of a thought by the conversants. This phenomena has social as well as psycholinguistic implications. .Bernstein (1956, p0 315) states that usually the sympathetic circularity phrase is said in response to phenomena outside the experience of the social order, or something peculiar to the social group. The circularity discourages reflexion on or consideration of any idea outside of the repeated phrase, and assures that the response to the idea will be expressed within the reasons or limits functioning in the restricted code. Bernstein feels that by using this code one has effectively limited curiosity and strengthens the solidarity of the social relationship. The frequent use of traditional phrases or the idiom, counters Bernstein, effectively increases the unity among a social group in the following way: Instead Of finding a way to express his feeling in a unique manner, the expres— sion of ideas by the use of the idiom serves as a referent in that others of his social order express their feeling in the same manner. Although the speaker may have had the Option to create language to express his individual feeling, he chose instead to use the traditional sayings, thus minimizing curiosity and creativeness. l4 Bernstein considers the last characteristic the most important variable and perhaps the product of the other nine variables. He states that a restricted code user's language is one of implicit meaning. This is a very salient point socially, psychologically and linguistically, because what is not said, may be just as meaningful and important as what is said. Linguistically, this would tend to isolate dialect groups because only members of the social order could inter- pret the code, i.e., interpret what was 395 said, which in fact was meaningful to the greater content of what was ex— pressed. Socially the code makes this group an identifiable social order, and for those who cannot switch codes psycho— logically it restricts their ability to express a unique idea to the confines of the code. In support of this concept, Erickson (1969) reported that the language used by the black speakers was adequate for communication of abstraction when the speakers shared the same linguistic context. He further notes that neither of his identifiable social groups were bound exclusively to any particular language code. Bernstein emphasizes that a restricted language code may be adequate for expression Of abstraction but the degree of specificity is relegated to a low level. The following example is used to exemplify the concept of degree-Of specificity. (A) Three boys are playing football and one boy kicks the ball and it goes through the window--and the boys are looking at it-—and a man comes out and 15 shouts at them--because they've broken the window-- and then that lady looks out of her window—-and she tells the boys off. (B) They're playing football and he kicks it and it goes through there——it breaks the window and they're looking at it-—and he comes out and shouts at them-— because they've broken it--so they run away--and then she looks out and tells them Off. Bernstein would suggest that the former is an elaborated code and the latter is the restricted code. The former ex- ample is less context bound than the latter, i.e., it is more difficult for an encoder to fully understand what is being said: the decoder need not have seen the actual event. The latter, then, depends upon what Bernstein terms universalis— tic knowledge about the situation to derive meaning from the discourse. Bernstein conducted two major research projects to sup— port his theory (Bernstein, 1962a, and Bernstein, 1962b). Bernstein used several of the same subjects in both of his studies. The subjects were sixty-one Caucasian males between the ages of fifteen and eighteen from a London public school. Bernstein chose the topic "The Abolishment of Capital Punishment" as his controlled stimulus tOpic. A magnetic tape recording sample of 1800 words which followed the first five minutes of discussion was selected for his analysis. In his first study (1962a), Bernstein measured mean phrase length, mean word length, and mean number of pauses between phrases (hesitation phenomena). He found socio-economic class to be a significant variable in identifying those who 16 used a "restricted" code as Opposed to those who used an "elaborated" code. Bernstein concluded from this study that there is a difference between working-class and middle class linguistic codes which are revealed in verbal planning procedures as measured by hesitation phenomena. In his second study, Bernstein (1962b) employed the language samples Of ten of the original forty-five middle class Caucasian boys and fourteen of the original language samples of the working class boys. Using his nine linguistic measures, Bernstein found support for his hypothesis that social class is a significant variable in the selection of language codes. Bernstein (1962) put forth the following discriminating hypothesis as indicative or predictive of social economic status as it interacts with the language codes. E1: Workinginlass families will exhibit a greater proportion Of pronouns/total words than will middle class families. 32: Working class families will exhibit a greater prOportion of pronouns 'you' and 'they' to total pronouns than will middle class families. H3: Working class families will exhibit a greater prOportion of sociO-centric sequences/total words than will middle class families. £4: Middle class families will exhibit a higher propor- tion of the pronoun 'I'/tota1 pronouns than will working class families. H5: Middle class families will exhibit a higher prepor- tion of ego—centric sequences/total words than will working class families. 17 In: a: Middle class families will exhibit a higher pro— portion of subordinate clauses/total finite verbs than will working class families. H7: Middle class families will exhibit a greater propor— tion of passive verbs/total finite verbs than will working class families. Im m Middle class families will exhibit a higher propor- tion Of the preposition 'of'/'of' plus ‘in' plus 'into' than will the working class families. The operational variables used by Bernstein to measure the above-hypotheses are as follows (the specific examples are adapted from Petersen, 1971): Hypothesis I: Pronouns/total words are essentially a type of.noun that cannot take the word 'the' immediately infront of it. Doubtful words are tested by trying to insert a 'the' in a sentence similar to the one in question. Ex: is very sad. went away. Examples of pronouns are I, you, he, she, it, we, they, me, yours, him, her, its, us, them, mine, yourself, his, hers, itself, ours, theirs, none, anyone, nothing, one, two, few, many, either. Hypothesis II: The pronouns 'you' and 'they' per total pronouns were defined by counting each instance of use of the pronoun by the subjects. Hypothesis III: Examples of socio—centric sequences (S.C.)/total words are: 'isn't it,‘ 'wouldn't it,‘ 'wouldn't you,’ 'ain't it,‘ 'you know.‘ Socio—centric sequences are usually used at the end of a sentence. Bernstein suggests the "sociO-centric sequences which are generated basically by uncertainty, maybe transmitted as a response of the speaker to condensation of his own meanings. The speaker requires assurance that the mes- sage has been received and the listener requires an opportunity to indicate the contrary.... S. C. sequences test the range of identifications which the speakers have in common." (1967, p. 205) Hypothesis IV: The pronoun 'I'/total pronouns is de- fined simply by counting each instance Of use of the pronoun in a subject's speech. _18 Hypothesis V: Examples of ego—centric sequences (E.G.)/ total words are: 'I think', 'I believe', 'I know', 'I mean,‘ etc. As with socio-centric sequences, E. G. sequences are generally found at the end of a sentence, although sometimes at the beginning when the verb Of the sequence is not the main verb of the clause. E. G. se- quences 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 related 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 E. C. sequences used more frequently by elaborated code users, and S. C. 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 VI: Subordinate clauses/finite verb or dependent clauses are introduced by subordinate conjunc- tions such as: as, as if, because, before, if, since, that, till, unless, when, where, whether. According to Walsh, a subordinate clause does not make sense when standing alone, that is, it is dependent upon other words to give it 'complete' meaning. The subordinate conjunc- tion connects two clauses of unequal rank, a dependent clause and an independent clause on which it depends. Examples of subordinate clauses are: 1. I was here before you came. 2. Robert fell in the water before he came to school. 3. When I leave this thesis Willibé‘COmpleted. Hypothesis VII: Passive verb/finite verb 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 auxillary verb 'be' followed by a past participle. Examples of passive verbs are: l. The man was hit by the car. 2. The mother had been worried by her son Hypothesis VIII: The preposition 'of,‘ in proportion to the use of.prepositions 'Of' plus 'in' plus 'into' were counted by noting each instance they occurred in the subject's speech. Lawton (1968) found support for Bernstein's restricted and elaborated code classification in his study which in- volved written as well as verbal discourse. Lawton limited 19 his sampling population to four groups, each of which con- tained five-boys. Utilizing Bernstein's nine linguistic measures, Lawton concluded that socio-economic class was a significant variable in the selection of a "restricted" verSUS an ”elaborated" code in both written and spoken language samples. .RObinson's (1965) investigation examined the possibility that the use of a relatively "restricted“ code by working- class boys was a matter of preference rather than inability to speak in a more "elaborated" code. Using two forms of written discourse (a formal and an informal letter) as his sampling method, RObinson found partial support for Bernstein's hypothesis. Robinson utilized 120 girls and boys between the ages of twelve and thirteen in his study. However, Robinson failed to control the tOpic of discussion. Robinson also utilized Taylor's cloze procedure (1953), since a portion of Bernstein's theory relates to the degree to which the lexical items Of a given language sample are predictable. Robinson concluded that social-class differences did not appear when the pressure on the working-class and middle-class groups was to use an "elaborated" code in a formal letter. Robinson did find differences in the language codes when he analyzed the informal.letters of his subjects. His latter analysis of informal letters were generally similar to those previously found by Bernstein (1962b) and Lawton (1968). 20 Poole and Field (1972) found support for the Bernstein theory. Their linguistic analysis, however, excluded the cumbersome analysis of lexical items. They utilized such measures as the abstract index, cloze procedure, and hesita— tion phenomena. Poole and Field used 10 males and 10 females in their study. Their limited pOpulation, however, limits their generalizations. McLaren (1968) cautions against universalization of Bernstein's theory, specifically in Australian society. He makes this statement on the basis that different class struc- ture and values in one country may be different from those of Britain. (Bernstein's theory is based heavily upon social interaction.) In American society, for example, little girls are rewarded for elaborate verbalization. The female in American society receives dolls and tea sets which encourage verbalization, whereas, little boys receive trucks and airs planes. The male who is talkative in American society is often thought of as affeminate. This male—female dichotomy would lead one to suspect that females would display language behavior more similar to an elaborated linguistic code and males would display language more indicative of restricted language code in American society. Petersen (1971), when considering the ten variables indicative of Bernstein's restricted and elaborated lin- guistic codes, found support for four of the linguistic variables, partial support for three of the variables and no 21 support for the remaining three variables. Petersen's sampling population consisted of thirty-two subjects. Included in his pOpulation were sixteen white males between the age of fourteen and sixteen years and their natural, mothers. Petersen also found support for another facet of Bernstein's theory which says that the teenage boys will exhibit linguistic behavior more like his mother's than his peer's. When reviewing the research in support of the Bernstein theory, it becomes apparent that all but Robinson (1965) and Poole and Field (1972) used Caucasian male subjects and generalized their theoretical finding to the linguistic behavior of pOpulations which include peOple of different racial and ethnic backgrounds. Their interpretations were also generalized to include both sexes. To date this researcher has been unable to find any empirical data to support such a generalization. In the field of education, child psychology, and socio- linguistics in America, there has evolved a theoretical language model which uses 'Standard American English' as a reference level. (Keep in mind that Bernstein was comparing Queen's English with Cockney.) Using standard American English as a reference, these theorists assumed that any linguistic form which shares American English is a pigeonized version. These theorists have become known as supporters of a language deficit hypothesis relative to ”non—standard" speakers of American English. 22 In Opposition to the language deficit theorists are those who suggest that because one's linguistic form shares some of the same phonological, grammatical, and semantic rules as Standard American English, this does not negate the fact that this linguistic form, because of its consistent rules of language, might be termed a different language. These latter theorists have become known as supporters of a difference-only hypothesis relative to "non-standard" speakers of American English. When reviewing the research on the socio-cultural aspects of poverty published during the mid and latter nineteen sixties, one is immediately struck with the apparent coalescence of interpretation about the language of the dis- advantaged urban Black American child. The Black child's speech has been labeled non-standard Black dialect. The language deficit theorists assume that the speaker of non- standard Black dialect speaks a form of sub-standard American English. According to the difference-only theorist, the non— standard Black English is a different language which merely happens to share some of the grammatical, semantic, and phonological rules with Standard American English. The language deficit theory emerged from the research and discussions of Deutsch and associates (1965), Bereiter and Englemann (1966), Hess and Shipman (1965), and Raph (1965). Throughout such literature, references are made to restrictions of the choice of lexical Options, the redun- dancies and inadequacies of the Black child's linguistic 23 system to convey abstract ideas. Such references are given theoretical support by citing Bernstein's sociolinguistic theory and associated speculations. While few of the deficit theorists explicitly equate their language paradigm with Bernstein's, they imply a similarity through the use of analogies. Hess and Shipman (1965) for example, incorporated Bernstein's restricted and elaborated codes into their research design and actually cite Bernstein in their review of the literature. The question Hess and Shipman were attempting to answer was: "What is cultural deprivation and how does it act to shape and depress the resources of the human mind?" The following is an example of how Bernstein's model was implicitly used by Hess and Shipman to describe the linguistic behavior of lower—class Black children. These styles of cognitive behavior are related, in our hypothesis, to the dimensions of maternal linguistic codes and types of family control systems. A status- oriented statement, for example tends to offer a set of regulations and rules for conduct and interaction that is based on arbitrary decisions rather than upon logical consequences which result from selection from one or another alternatives. Elaborated and more per- son-oriented statements lend themselves more easily to styles of cognitive approach that involve reflective comparison. Status-oriented statements tend to be restrictive in thought. Several compensatory language and reading programs de- velOped for the culturally disadvantaged youth in Urban America are based upon the Bernstein dichotomy of restricted and elaborated language codes. Bereiter and Englemann, for example, equated "language deprivation" with educational deprivation. Their linguistically deprived model paralleled 24 that of the Bernstein restricted code. Horn (1970) in his reading program for the culturally disadvantaged Black child described the child in his introductory chapter in terms Of the restricted-elaborated code dichotomy. The compensatory education program and teaching strategies prOposed by Taba and Elkin (1966) is also based upon the descriptions and’ hypothesis of Bernstein's model. Summary and Statement of the Problem In summary, the Bernstein language model of restricted and elaborated language codes has received support by the empirical research of Bernstein (1962a and 1962b), Petersen (1970), Poole and Field (1972), Lawton (1965), and Robinson (1968). However, these researchers limited their data to the significance of socio—economic class, as relavant to selec- tion of either language code. The populations in these studies are characterized as Caucasian male, except for the Robinson and the Poole and Field's studies. American re- searchers, educators and psychologist (Deutsch, 1965; Bereiter and Englemann, 1966; Hess and Shipman, 1965; Raph, 1965; Horn, 1970; and Taba and Elkins, 1966) have suggested that the Black ghetto language is a substandard version of Standard American English. They utilize the restricted— elaborated paradigm to denote that the lower-class, Black child is speaking a substandard form of English. Bernstein (1970) Objects to such a use of his language codes. 25 Furthermore, McLaren (1968) Objects to the transfer of the Bernstein model on the grounds that Australian society, with its different class structure and values, could not be compared to British society sociolinguistically using this model. The present investigation makes the same argument for speakers of non—standard Black English. Bernstein com- pared the Queen's English with Cockney; the Americans are comparing non-standard Black English with Standard American English. In the American interpretation of Bernstein, the promoters of the American compensatory education programs, child psychologists, and others generalize Bernstein's theory to include both males and females, Black children as well as White children. To date this researcher has not been able to find empirical research to warrant such a uni— versalization of elaborated or restricted language codes. An investigation was designed to answer the,following questions: 1. Is sex a significant variable in the selection of a restricted or elaborated language code in American society? Will American males and females receive similar scores on the eight linguistic measures used by Bernstein? 2. Is socio—economic status in American society a significant variable in selection of a restricted or elaborated language code? Will upper middle- class, lower middle—class, upper working-class and lower working—class subjects receive similar rating on the eight linguistic measures? 3. Is race a significant variable in the selection of a restricted or elaborated language code? Will Black Americans and White Americans display similar ratings on the eight linguistic measures? 26 In general, then, this study attempted to answer the ques— tion: Does race, sex, or socio-economic status interact to produce significant differences in ratings on the eight linguistic variables? Further, do they influence the selec— tion Of a restricted or elaborated language code? CHAPTER II EXPERDMENTAL PROCEDURES This study utilized forty tape recorded language sam- ples of Black and White male and female teenagers. These language samples were collected from a study carried out by Shuy, Wolfram, and Riley, Field Techniques nggp Urban Language Study (1968). Sociometric measures were performed by these investigators on these forty subjects. Transcripts were made of the forty tapes. Using eight of the linguistic measures described by Bernstein (1962a), a linguistic analy- sis was performed on these samples by a bidialectic Black male analyzer. A leastsrsguares solution to an analysis of variance with unequal N's was employed upon each of the eight measures. Graphs were constructed to depict the corre- lation between race, sex, and socio-economic status. Subjects The speakers for this study were tape recorded language samples (Shuy, Wolfram, and Riley, 1968) of fortyrteenagers between the ages of 13 and 18 years old from Detroit, Michigan. The sample consisted of IS Black males, 7 White males, 9 Black females and 11 White females. Thirteen of the S's were 27 28 from an upper middle-class socio-economic level; six were from a lower middle-class sociO-economic level; twelve were identified with the upper working-class socio-economic level; and twelve were identified with the lower working-class socio—economic level. The speakers were classified by Shuy £5 21., who utilized Hollinghead's (1958) scale of social stratification. It combines scales of education, occupation and residency in computing the relative social rank of informants. Table I relates the social class designation to the social status scores.l Figure 1 reveals the social class distribution of the black and white subjects used in this study. For purposes of having as many subjects in each cell as possible during the statistical treatment, lower and upper working-class categories were collapsed together and upper and lower middle-class categories were considered as one socio-economic status. The interviews had been conducted in the homes of each subject. The interviewers were white graduate students in the Linguistics Department at Michigan State University. The interviewers confined the formal portion Of the interviews to the following topics: 1. Recent television programs 2. Ways of obtaining a wish, 3C Explanations of their favorite games 4. Descriptions of their school and playground 5. Discussion of their favorite teacher 6. Discussion of their worst teacher The tape recorded interviews were transcribed by a bi- dialectic Black male graduate student in Speech pathology, 29 proficient in the use of non-standard Black dialect and Standard American English. Table 1. Distribution of social status scores in relation- ship to the four social class labels. Class Designation Social Status Score Upper middle-class . 20-48 Lower middle-class 48-77 Upper working-class 78-106 Lower working—class 107-134 Transcription Procedures The transcriber, utilizing a Wollensak T-1500 tape recorder with a foot Operated pause switch, transcribed ver— batum all of the conversation on the forty tapes. The words of the interviewer were written in red ink and the words of the child being interviewed were written in blue ink. The experimenter timed two consecutive five minute seg- - ments of the interview. The five minute segments were marked on thevtranscript. If the five minutes elapsed during a portion Of a clause, the experimenter included the complete fclause within the five minute time segment. 30 100 4 ----- Black --—-White Percent of Total Population Upper- Lower- Upper- Lower- middle middle working working Figure l. The percentage of distribution of Black and White subjects throughout the four socio—economic levels. 31 Linguistic Analysis The experimenter utilized the following measures out- lined by Bernstein (l962a) in the analysis of the forty transcripts: I. Pronouns per total words II. Pronouns "you" and "they" per total pronouns III. Socio—centric sequences per total words IV. Pronouns "I” per total pronouns V. Ego-centric sequences per total words VI. Subordinate clauses per total finite verbs VII. Passive verbs per total finite verbs VIII. Number "of's" per total "of," "in," and "into" A coding system was devised which inabled the researcher to re-evaluate any word or phrase which he had previously labelled as a pronoun, finite verb, subordinate clause, etc. Each linguistic measurement was assigned a number from one to eight with the exception of the finite verbs, which were labelled 'F'. As the researcher analyzed each word or phrase in terms of its grammatical set, the code number or grammati- cal set under which it was considered was placed immediately above the word or set of words. Since individual words, theoretically, could be considered members of more than one set or measurement, more than one code number could appear above a given word. An actual linguistic analysis would appear as follows: 1 4,5 l 2 F‘ l I hfard Ehat you break the bone and this person be one person hold one end of the bone. Only one linguistic measurement was considered at a time. The researcher analyzed all forty transcripts in terms of one 32 measurement. The total was obtained and recorded under the code and labelled on the outside jacket Of each transcript. The total score for each five minutes of discourse appeared on the outside jacket as a ratio. These two five minute ratio scores were combined to obtain a total ratio score for a ten minute segment, i.e., pronouns per total words = 256/1256. The total ratio was computed and trans- formed into a percentage score. For example, a ratio of 256/1256 was equal to 20.3%. CHAPTER III RESULTS This thesis investigated the feasibility of applying Bernstein's ”restricted" and "elaborated" language code model to American pOpulations. The group under study consisted of Black and White male and female teenagers from both middle and working—class socio-economic groups. Eight of the ten measures Bernstein (1962) reported indicative of a "restricted code" or an "elaborated code" were used to analyze the forty language samples. The eight measures were as follows: Pronouns/total words, Pronouns "you" and "they”/total pro— nouns,. Socio-centric sequences/total words, Pronouns "I“/ total pronouns, Ego-centric sequences/total words, Subordinate clauses/total finite verbs, Passive verbs/total finite verbs, and Total "Of's"/tota1 "of" "in," and "into." For each of the eight measures used in this study a leasts squares soltuion to an analysis of variance for un- equal N's was employed to reveal possible significant effects associated with each of the main factors of race, sex, and socio—economic level, and their respective interactions. In addition, an intercorrelation matrix between the eight measures was computed. Significant effects were associated 33 34 with the following measures: Pronouns "you" and "they“/total pronouns, Ego-centric sequences/total words, subordinate Clauses/total finite verbs, and Prepositions "of'I "of" "in" and "into." These four measures are reported upon below, and can be found illustrated in Figures 2 through 6. The re- maining four measures, Pronoun/total words, SociO-centric sequences/total words, Pronouns "I"/total pronouns, Passive verbs/total finite verbs, had no significant effects associ- ated with them. Effects Associated with Race Only two of the several measures used showed significant effects for the factor of Race (Black versus White). There was a significant main effect (F = 8.20, p-<.007) for the measure of Subordinate clauses/total finite verbs. Figure 2 indicates that Black Americans tended to use more subordinate clauses/finite verb, than white Americans in this population. This finding is incongruous with the current version of Bernstein's model as interpreted and applied by American edu- cators and psychologists. There was a significant interaction between Race and Sex (Black males versus White males versus Black females versus White females) for the measurement Preposition "of/"of" "in" and "into" (F = 4.83, p<:.04). This interaction is illus- trated in Figure 3. According to the Bernstein hypothesis, middle-class speakers will exhibit a higher score on this 35 measure, indicative of the use of an elaborated code. As can be seen in Figure 1, there was a larger distribution of Blacks in the lower socio-economic levels of this popula- tion. As illustrated in Figure 4, Black males tended to use more "Of's"/total "of" "in" and "into," than Black females, White males, or White females. White females attained the next highest score on this measure, followed by White males, Black females, respectively. These results, though they appear mixed, did not follow the predicted trend of the Bernstein language model. Effects Associated with Socio-economic Status Only two of the measures used showed significant effects associated with socio-economic status (working-class versus middle-class). There was a significant main effect for socio- economic status (F = 10.95 p‘<.002) on the measure of Subordi— nate clauses/finite verbs as can be seen in Figure 4. The subordinate clause, according to Bernstein, is a more sOphis— ticated means of introducing an idea. Thus, the percentage of use of this measure is also directly proportional to the degree of “elaborateness.” As predicted by the Bernstein hypothesis, middle-class speakers score higher on this measure than working-class speakers. There was a significant interaction between socio-economic status and sex (middle-class males versus middle-class females versus lower-class males versus lower—class females) on the 36 measure of Ego-centric sequences/total words (F = 4.1, p<:.05). As illustrated in Figure 5, middle-class females and working-class males both made similar high scores. In contrast to this, middle—class males and working-class females made identically lower scores. In keeping with the restricted versus elaborated dichotomy, one who tends to use a higher percentage of ego—centric sequences is regarded as a restricted language code user. Effects Associated with Sex Three of the several measures used showed significant effects for the factor of sex (males versus females). There was a significant main effect (F = 12.3, p1<.001) for the measurement of Pronouns "you" and "they"/total pronouns. Figure 6 indicates that female subjects tended to use more "you's" and "they's"/total pronouns than male subjects. Bernstein would suggest that frequent use of "you" and "they" is indicative of a restricted language code. As noted previously and illustrated in Figures 3 and 5, there appeared a significant interaction between sex and race and also sex and sociO-economic status on the measures "of"/tOtal "Of," "in," and "into," and ego-centric sequences/ total words, respectively. It is evident from the findings reported in this chapter that seven of the null hypotheses postulated by Bernstein (1962) and Petersen (1970) must be rejected. The single 37 hypothesis supported by the present investigation was that middle-class speakers tended to use more subordinate clauses/ total finite verbs. Intercorrelations of the Measures As illustrated in Table 2, there were significant corre- lations between three pairs of the several measures used. Measure 1, pronouns/total words, showed a significant positive correlation with measure three, socio—centric sequences/ total words. The socio—centric sequence as explained in Chapter I usually appears at the conciusdon of a sentence. The sequence consists of a pronoun and a verb,i.e., 'ain't it,‘ 'you know'. Thus it can be seen that the measure pronouns/ total words can be viewed as a subset of the measure socio- centric sequences/total words; therefore, it is quite logical that everytime the experimenter counts socio-centric sequences he also is considering pronouns. The less than perfect posi- tive correlation is related simply to the fact that not all pronouns were Observed in conjunction with a socio-centric sequence. Measure 4, pronoun "I"/tota1 pronouns and measure 5, ego—centric sequences/total pronouns, exhibited a significant positive correlation as can be seen in Table 2. The ego- centric sequence as explained previously consists of the pronoun "I" plus a verb, i.e., "I think", "I mean" or "I believe." The measure pronouns "I"/total pronouns can be 38 considered a subset Of the measure ego—centric sequences/total words because the elements contained in measure 5 are also present in measure 4. The less than perfect positive corre- lation can be attributed to the observation that not all "I's" Observed in the sample are in the context of an ego-centric sequence. The correlation matrix also illustrated a significant negative correlation between the use of measure 2, pronouns "you“ and "they"/tota1 pronouns, and measure 4, pronoun "I"/total pronouns. This can be attributed to the possible lexical options Open to any speaker using pronouns. The encoder using pronouns either has to talk about himself, wherein he uses the pronoun I, or he has the alternative Of expressing an idea about someone or something other than .himself. The reference to an 'other' is formulated linguis- tically through the use of "you" or "they." There are other possible lexical options such as using nouns like boy, girl, etc. The less than perfect negative correlation is undoubt- edly related to the experimental subjects' use of the alternative lexical options such as he, JOhn, she, etc. 39 [2% I u: I 10% I” 8% v 72 v- 6% 1" sz .. 22 4- n: .- 02’ Black 6 Americans Americans Figure 2. Mean percentage of subordinate clauses per total finite verb as a function of race. 39 (2% I IIX 1* I02 1* 9% «I 8% v- 7; 1*- oz. 'I— SEZ T 4211" ;3; .. 122 ~- (OZ 6 Americans Americans Black Figure 2. Mean percentage of subordinate clauses per total finite verb as a function of race. 4O 60%: T 552“ 5311 40%: J I 351 r 3021 2531 T I 2034 152 ~~ [03 ~ T 451-- £92 Black Black White White Males Females Males Females Figure 3. Mean percentage of preposition "of" per "of" "in" and "into" as a function of Race as it interacts with Sex. 41 12% 1: “7. v [0% '- 91: 1~ 8% 't 6% -+ 52 I 22:1P I’l. 1- 07: Middle Working Class Class Figure 4. Mean percentage of subordinate clauses per total finite verbs as a function of SES. 42 [0% ‘- 31 .. ‘7‘ 4.. .63‘v .42 w .22. ~- Middle Middle Working Working Class Class Class Class Males Females Males Females Figure 5. Mean percentage of ego—centric sequences per total words as a function of Sex as it interacts with SES. 43 45; ~~ 4a; h 303 I 202 .. 15‘: -- to: +1: 02 Males Females Figure 6. Mean percentage of pronouns "you" and “they" per total pronouns as a function of Sex. 44 ooooo.a mmmmm.o mvham.o omooo.o hmo¢0.0 dammm.o NMNmH.OI HmHOH.OI m.mmoz ooooo.a oaoem.o mmmmo.o mmoeo.o mamaa.o omamo.on mo¢wm.o h.mmo2 ooooo.a mmonm.01 enamo.o mmamo.on mmmmo.o: m.mmo2 OOOOO.H som¢N¢.o mmmho.o mHomN.OI OHNGH.O m.mm02 ooooo.H dehmo.o «vvvmm.01 Newmo.o ¢.mmm2 OOOOO.H mbbma.o sdmmmv.o m.mm02 ooooood mhmHN.O N.mmm2 ooooo.a H.mmm2 m.mmo2 >.mmoz m.mooz m.mmo2 ¢.mmoz m.mmoz. N.mmoz H.mmoz .uopoo poumuonmao pas pouowuymou mo mousmmoe swoumcuom names How xfluuma coflumaouuou .N manna CHAPTER IV DISCUSSION The present study investigated the Bernstein elaborated and restricted language code model as applied to an American pOpulation. The population under study consisted of Black males, Black females, White males, and White females. All of the subjects were teenagers from working-class and middle- class families living in the Detroit metropolitan area. The language samples collected from these teenagers were obtained during the Detroit Negro Dialect Study. Bernstein (1962) set forth eight linguistic measures which he stated were discriminating between restricted and elaborated language codes. This research related these eight measures to the factors of race, sex, and socio—economic status. At best, only four of the eight measures are capable of discriminating "elaborated" from so-called "restricted" code usage. When considering these four discriminating measures, only two are related to race and both of those are not independent measures, but rather are dependent upon other factors (sex and socio-economic status). This investigation utilized the same measures as Peter- sen (1970), Lawton (1968) and Bernstein (1962). The findings 45 46 in this research did not support either of its predecessors. Though the major emphasis in these studies was socio- economic status, Petersen's research and Poole and Fields' research were the only investigation with American pOpula— tions. Petersen's pOpulation consisted of sixteen teenagers from the various sociO-economic classes, which may have put serious limitations upon the generalizations he could make about his results. In addition, Petersen considered the person from a family Of four who earns $8,000 dollars per year and who lives in Holt, Michigan as lower-class. Granted that there are considerations other than income associated with a sociO-economic status classification, this particular practice is Obviously open to question. A portion of Petersen's support for Bernstein's hypothe- sis was based upon Gillies abstract index (1958) which measures the degree of "abstractness" of a written piece Of material. Petersen in his dissertation sights Robinson (1965), as one who utilized the index to investigate Bernstein's theory. .Robinson specifically had his subjects write letters under a formal condition and an informal condition. These written letters were then subjected to the procedures neces- sary to determine the abstract index. .However, Petersen took spoken discourse, transcribed it from a magnetic tape record— ing into writing and then subjected it to the procedures necessary to obtain an abstract index. By doing this, 47 Petersen assumed that the grammatical sets used during spon— taneous conversation are equivalent to those used during written discourse. During spoken discourse, however, the encoder has the decoder present in most cases. This is a salient feature because, as the encoder speaks, he receives feedback from the decoder concerning the decoder's under- standing of what is being said. The feedback that the encoder receives will determine the structure and perhaps even degree of "specificity" involved in his message. In contrast to the above, the person who writes a message has the distinct disadvantage of not having the decoder present. Therefore, he does not receiVe any feedback. In his writing, the encoder must anticipate and infer certain things about the reactions of the decoder; therefore, written discourse may of necessity be more detailed. Thus one must seriously question the practice of equating written discourse with spoken discourse as Petersen had done in his research. In support of this same idea, Petersen using his American population found no support for the hypothesis associating a higher abstract index with maddle-class speakers, and only a partial support for the remaining hypothesis. Petersen concedes in his discussion that he may have been in error by incorporating the abstract index to measure the degree of abstraction with spoken conversation. The findings in this study on Americans is also incon- gruous with the conclusions reached by Poole and Fields (1972). 48 However, as pointed out in Chapter I, Poole and Field did not perform an analysis of the lexical items. They chose the measures of abstract index, cloze procedure, and hesitation phenomena. Poole and Fields' techniques measured the "restrictiveness” or "elaborateness" of a language code through indirect means. Bernstein conceptualized that the lexical options of a restricted code speaker are more limited and therefore are more predictable, and it is on this basis that Poole and Fields rationalized using the cloze procedure. It is also reported that hesitation phenomena is directly related to the verbal planning stages of speech. Thus, they assumed that since elaborated code speakers exhibited a higher score on this measure their linguistic output was more "complex," or at least required more "complex" preverbal planning. The correlation between these two independent variables (grammatical complexity of linguistic output and hesitation score) was assumed valid by these experimenters, though the American pOpulation's stress and inflexional pat— terns are markedly different from the‘”Queen's English." Bernstein predicted that middle-class speakers will tend to converse in more abstract terms. Poole and Fields chose the abstract index to measure this phenomena. It is quite apparent from this brief discussion of Poole and Fields' research that they incorporated completely different measures than were used in this investigation, and originally suggested by Bernstein. These two studies, then, 49 cannot be compared even though they look at the same theory from different perspectives. It is interesting, however, that by having these discrepancies, Poole and Fields' conclu- sions were the anti—thesis of the findings in this investiga- tion. One of the paramount questions raised by this investi- gation was "Could Bernstein's theoretical language codes be applied to differential Black and White male and female speakers in America?" The answer to this question, in view of the present investigation, is negative. Further, Bernstein's measures do not correlate with the main factor of socio- economic status in American subjects, except on measures 5 and 6, one of which is dependent upon the interaction Of race. As mentioned before, the only measure which supported Bern- stein on socio-economic status was variable six (subordinate clauses/total finite verbs). In conjunction with these findings, it may be that the language patterns spoken by the working-class Black subjects and a few of the middle-class Black subjects contributed to the rejection of the theory. The language patterns used by the White subjects were more similar to the Queen's English (upon which Bernstein's research is based) than the "non- 'standard Black dialect" spoken by many of the Black subjects. It would appear from these results that the use of the "restricted" and "elaborated" labelling system to describe children who speak "Black dialect" appears to be erroneous, 50 in addition, to its application to American pOpulation in general. Most of the investigations of Bernstein's theory have had a tendency to use small populations consisting roughly of less than twenty subjects. This is especially true for those studies that incorporated linguistic measures involving analysis of lexical items. Poole and Fields state in their discussion that a linguistic analysis of the lexical items is cumbersome. This is a pOint well taken. However, one cannot make gross generalizations about the linguistic be- havior of a large pOpulation without examining in depth a sufficient number of peOple from that pOpulation. Though the present investigation inCorporated the largest N to date Of any linguistic study of restricted versus elaborated codes using these eight linguistic measures, it is nonetheless felt that generalizing about the linguistic system of the entire pOpulation of the Detroit metropolitan area based upon forty speakers may be Open to question. However, speculations without any empirical research was equally unwise, which is what had been done up until this time. Implications for Future Research Several of the language samples selected from the Detroit Negro Dialect had been transcribed prior to this researcher's acquisition of the language tapes. To control for differences 51 between transcribers, it was decided that all of the tapes were to be retranscribed by the present researcher, a Black male native speaker of "Black dialect" and "Standard American English." Upon inspection of the transcriptions which had been completed by the White male and female linguists, it was noted that there was an abnormally high frequency Of semantic and grammatical errors in the transcriptions Of lower-class Black informants. This raises the following questions: 1. Were there any consistencies in the specific types of errors (vowels, substitutions, omission, etc.) committed by these linguists? 2. Over time did the'frequency of errors decrease as a product of familiarity with the dialect? 3. What specific semantic, syntactic, and phonological rules did these linguists lack which were necessary to evaluate a particular response when it appeared distorted on the tape recording? One of the major problems associated with most socio- linguistic studies is that of obtaining a language sample which is truly representative of the informant's linguistic system. As has been pointed out in experimenter bias studies by such investigators as Rosenthal and Jacobsen (1965) the researcher or person collecting the data has almost as much influence upon the quantity and quality of an experimental sample collected as does the person who is giving the sample. This experimenter effect may have nothing to do with the I. experimenter's actions per se. The informant may react or report or speak in a particular fashion because of his 52 perception Of the social set to which the experimenter belongs. Labov (1970) in a taped interview discusses the problems associated with having a White male or female interview a lower-class Black child from an urban area. In summary, he states that regardless of what the White experimenter says he is usually regarded as and perceived as one who is different, and in most ways an insignificant other; therefore, the language he uses may reflect the presence of this alien. Thus the questions are raised: What are the experimenter effects associated with the sex Of the experimenter? What are the effects associated with the differences between the race and social class of the experimental subjects? In light of the findings of this investigation, it would appear that this is but one step in a series of investiga- tions which should be designed to empirically investigate the underlying theoretical assumption made by child psychol- ogists such as Conger, Kagan and Mussen (1965), who discuss the Bernstein theory as applied to Black and lower-class people as though it was fact. What is even more disturbing is that their introductory text to child psychology is used in many major universities and teacher—training institutions in America. This investigation found that during the transcription of the tapes an intimate knowledge of the subtleties of Black dialect was moSt important. It appears imperative that in future research involving peOple who speak Black dialect that 53 the investigator either speak the dialect or have someone interpret the dialect for him. CHAPTER V S UMMARY This thesis investigated the Bernstein hypothesis of restricted versus elaborated language codes as applied to an American pOpulation consisting of Black and White males and females from four major socio-economic groups. The forty language samples used in this study were Obtained from the Detroit Negro Dialect Study by Shuy, Wolfram and Riley (1968). Bernstein listed ten discriminating measures of a restricted and elaborated codes. Eight of these ten measures were employed in this research and are as follows: Pronouns/total words, Pronouns "you" and "they" per total pronouns, Socio—centric sequences/total words, Ego— centric sequences/total words, subordinate clauses/total finite verbs, Passive verbs/total finite verbs, Number "Of's"/ total "of" "in" and "into." These eight measures were correlated with the main effects of race, sex, and socio- economic status and possible significant interaction between these main effects. It was concluded from the results of this research that only four of the eight measures were capable of discriminating between a restricted code or an elaborated code. When con- sidering these four_discriminating measures, only two were 54 55 related to race, and both Of these are not independent measures, but rather are dependent upon the interaction of sex and socio-economic status. The results of this research showed that only one out of the eight hypotheses Bernstein formulated was verified; therefore, this research concluded that the Bernstein model Of a restricted language code versus an elaborated language code is not applicable to an American pOpulation consisting of Black and White males and females from various socio—economic classes. LIST OF REFERENCES Baratz, Joan C. "Teaching Reading In an Urban Negro School System," Language and Poverty, Frederick Williams, Merkham Press, Chicago, Illinois, pp. 11-23, 1970. Bereiter, Carl. ”The Relative Importance of Verbal and Non- verbal Factors in Cultural Deprivation: Evidence from Children with Sensory Handicaps." Urbana, Illinois: Institute for Research on Exceptional Children, 1965. , Englemann, Siegfried. Teaching Disadvantaged Children in Preschool, Englewood Cliffs, New Jersey: Prentice—Hall, 1966. Bernstein, Basil. ”A Public Language: Some Sociological Implications of a Linguistic Form," British Journal of Socioloqv. 10:311. 1959. . "A Sociolinguistic Approach to Socializations With Some Reference to Educability," Language and Poverty, Frederick Williams, Markham Press, Chicago, Illinois, pp. 25-58. . "Elaborated and Restricted Codes: An Outline," Sociological Inquiry, New York: 36 (2), 1966, p. 259. . "Linguistic Codes, Hesitation Phenomena and Intelligence," Language and Speech,£h Part 1 (OctOber- December 1962a), pp. 31-46. . "Social Class, Linguistic Codes and Grammatical Elements," Language and Speech, 5, Part 4 (OctOber- December l962a), pp. 221-240. . "Some Sociological Determinants of Perception," British Journal of Sociology, 1958, p. 161. , and Henderson, D. "Social Class Differences in the Relevance of Language to Socialization," Sociology, 3, (1), January, 1969. Deutsch, M. and Brown, B., "Social Influences in Negro— . White Intelligence Differences,“ Journal of Social Issues, 20:24-35, 1964. 56 57 Erickson, F., 'F' get you Honky?': A New Look at Black Dialect and the School, ElementarygEnglish, 46:495-517, 1969. Gillie, Paul J. "A Simplified Formula for Measuring Abstrac- tion in Writing,“ Journal of Applied Psychology, 41, No. 4: 214—218, 1957. Hess, Robert D. and Shipman, Virginia. "Early Experience and the Socialization Of Cognitive Modes in Children," Child Development, pp. 869-876. Labov, William. The Stugy of Nonstandard English, published by the National Council of Teachers Of English, 1970. . "The Logic of Nonstandard English," Language and Poverty, Frederick Williams, Markham Press, Chicago, Illinois, pp. 153-189. Lawton, Denis. Social Class, Language and Education, Routledge and Kean Paul, London, 1968. Petersen, Duane D. "A Socioloinguistic Study of Elaborated and Restricted Code Systems,” unpublished Ph.D. disser- tation,.Michigan State University, 1970. Piaget, Jean. The Language and Thought of the Child, New York: .Meridian Books, 1955. Poole, M. E. and Field, T. W., "Social Class and Code Elabora- tion In Oral Communication," Journal of Speech and Hearing Research. 14 (2): 421-427, 1972. Raph, Jane B. "Language and Speech Deficits in Culturally Disadvantaged Children: Implications for the Speech Clinician," Journal Of Speech and HearingyDisorders, 32: 203-214, 1967. .RObinson, W. P. "The Elaborated Code in Working Class Language," Language and Speech, Vol. 8, 1965. Rosenthal, R. and Jacobsen, L. "Self-fulfilling Prophecies in the Classroom: Teacher's Expectation as Unintended Determinants of Pupils Intellectual Competencies," in Deutsch §E_gl, ed. Social Class, Ragey and Psychological DevelOpment. New York: Holt, Rinehart, and Winston, 1968. Sapir, E. Culture, Language and Personality, University of California Press, 1956, p. 70. 58 Shuy, Roger, Wolfram, W. and Riley, W. Field Techniques in an Urban Language Study. Center for Applied Linguistics, _l970. . "Bonnie and Clude Tactics in English Teaching," The Florida F. L. Reporter, Vol. 7, Number 1, 1969. A paper presented at the 1968 NCTE Convention in Milwaukee. ' Taba, Hilda. “Cultural Deprivation as a Factor of School Learning," Merrill Palmer Quarterly, 10: 147-159, 1964. Taylor, W. "Close Procedure:_ A New Tool for Measuring Readability," Journalismyguarterly, 30: 415—433, 1953. Templin, Mildred. "Relation of Speech and Language Develop- ment to Intelligence and Socio-eOOnomic Status," Volta Review, 60: 33143347'1958. Whorf, B. L. Language, Thought and Reality, New York: Wiley and Sons, 1956. Williams, Frederick, Language and Poverty, Chicago, Illinois: Markham Press, 1970. , and Naremore, Rita.~ "On the Functional Analysis of Social Class Differences ianodes of Speech," Speech Monograph, 1969, 36:77-109. Wolfram, Walter. A Sociolinguistic Description of Detroig Negro Speech, Center for Applied Linguistics, Washington, D. C., 1970. APPENDIX A WEIGHTED AVERAGE SCORES FOR EACH MEASURE EMPLOYED AS A FUNCTION OF RACE, SEX, AND SES AND THEIR RESPECTIVE INTERACTIONS 59 WEIGHTED AVERAGE SCORES FOR EACHLMEASURE EMPLOYED AS A-FUNCTION OF RACE, SEX, AND SES AND THEIR RESPECTIVE INTERACTIONS WC MC WC MC Black 15.7 14.6 15.2 15.1 15.8 15.0 Measure 1 White 17.0 16.5 16.1 15.6 16.1 17.0 Total 32.7 31.1 31.3 30.7 31.9 '32.0 Black 33.8 17.4 18.1 17.1 21.6 18.2 Measure 2 WC MC WC MC Black .2 0.3 0.8 .6 .6 .0 Measure 3 White .8 .4 1.0 .5 .8 0.2 Total 1.0 .7 1.8 1.1 1.4 Black 18.7 21.3 18.8 22.2 19.1 22.2 White Total 25.7 59.5 27.4 44.8 15.2 33.3 24.2 41.3 21.0 42.6 28.4 46.6 Measure 4 White Total 18.9 37.6 20.1 41.4 28.1 46.9 19.5 41.7 24.2 43.3 19.7 31.9 continued 61 Measure 5 Measure 6 Black White Total Black White Total M 0.4 0.7 1.1 9.4 4.8 14.2 WC 0.7 0.7 1.4 7.0 8.8 15.8 MC 0.3 0.6 .9 12.2 11.7 23.9 F 0.5 0.5 1.0 9.6 6.0 15.6 WC 0.4 0.7 1.1 12.5 9.0 21.5 MC 0.7 0.5 1.2 1.5 13.0 14.5 Measure 7 .Measure 8 Black White Total Black White Total M 0.5 0.2 .7 37.6 45.4 83.0 WC 0.2 0.4 .6 51.6 42.4 94.0 MC 0.8 0.7 1.5 57.3 47.7 105.0 F 0.5 0.2 .7 52.3 46.5 98.8 WC 0.7 0.5 1.3 52.3 45.0 97.3 MC 0.3 0.6 .9 35.6 42.5 78.1 M = Male F = Female WC = Working class MC Middle class MICHIGAN STATE UNIV. LIBRQRIES 3129301 1039991